[{"_1":2,"_410":-5,"_411":-5},"loaderData",{"_3":4,"_22":23,"_278":279},"root",{"_5":6,"_7":8},"toasts",[],"visitor",{"_9":10,"_11":12},"viewportWidth",1280,"visitorSession",{"_13":14,"_16":17,"_18":-5,"_19":20,"_21":-5},"attributes",[15],"bot","id",19346472,"membership","referenceId","0199f420-7955-7299-8090-ea86aef37824","userAccount","routes/_public/mcp/servers/~namespace/~slug/_pages/_layout",{"_24":25,"_68":-5,"_69":37,"_27":70,"_194":195,"_277":176},"availableLocales",[26,37,42,47,53,58,63],{"_27":28,"_31":34,"_35":36},"mcpServer",{"_29":30,"_31":30,"_32":33},"description","mcp-victoriametrics","name","readme","# VictoriaMetrics MCP-Server\n\nDie Implementierung des [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) -Servers für [VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/) .\n\nDies ermöglicht den Zugriff auf Ihre VictoriaMetrics-Instanz und die nahtlose Integration mit den APIs und [der Dokumentation](https://docs.victoriametrics.com/) [von VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/url-examples/) . Es bietet Ihnen eine umfassende Schnittstelle für Überwachungs-, Beobachtungs- und Debugging-Aufgaben im Zusammenhang mit Ihren VictoriaMetrics-Instanzen und ermöglicht erweiterte Automatisierungs- und Interaktionsfunktionen für Ingenieure und Tools.\n\n## Merkmale\n\nMit diesem MCP-Server können Sie fast alle schreibgeschützten APIs von VictoriaMetrics verwenden, d. h. alle in [VMUI](https://docs.victoriametrics.com/#vmui) verfügbaren Funktionen:\n\n* Abfragen von Metriken und Erkunden von Daten (sogar Zeichnen von Diagrammen, wenn Ihr Client dies unterstützt)\n* Auflisten und Exportieren verfügbarer Metriken, Beschriftungen, Beschriftungswerte und ganzer Reihen\n* Analysieren Ihrer Alarmierungs- und Aufzeichnungsregeln und Alarme\n* Parameter Ihrer VictoriaMetrics-Instanz werden angezeigt\n* Untersuchen der Kardinalität Ihrer Daten- und Metriknutzungsstatistiken\n* Analysieren Ihrer Abfragen\n* Debuggen Ihrer Relabeling-Regeln, Downsampling- und Aufbewahrungsrichtlinienkonfigurationen\n\nDarüber hinaus enthält der MCP-Server eingebettete aktuelle Dokumentationen und kann diese ohne Online-Zugriff durchsuchen.\n\nWeitere Einzelheiten zu den genau verfügbaren Tools und Eingabeaufforderungen finden Sie im Abschnitt [„Verwendung“](#usage) .\n\nSie können die Funktionalität von Tools und Dokumenten in Ihren Eingabeaufforderungen kombinieren und großartige Anwendungsszenarien für Ihre VictoriaMetrics-Instanz entwickeln. Sehen Sie sich einfach den Abschnitt mit den [Dialogbeispielen](#dialog-example) an, um zu sehen, wie es funktionieren kann. Beachten Sie bitte, dass die Qualität des MCP-Servers und seiner Antworten stark von den Fähigkeiten Ihres Clients und der Qualität des verwendeten Modells abhängt.\n\nSie können den MCP-Server auch mit anderen MCP-Servern für Observability oder Dokumentensuche kombinieren und so noch leistungsfähigere Ergebnisse erzielen.\n\n## Anforderungen\n\n* [VictoriaMetrics-](https://docs.victoriametrics.com/victoriametrics/) Instanz ( [Einzelknoten](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/) oder [Cluster](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/) )\n* Go 1.24 oder höher (wenn Sie aus der Quelle erstellen möchten)\n\n## Installation\n\n### Gehen\n\n```bash\ngo install github.com/VictoriaMetrics-Community/mcp-victoriametrics/cmd/mcp-victoriametrics@latest\n```\n\n### Quellcode\n\n```bash\ngit clone https://github.com/VictoriaMetrics-Community/mcp-victoriametrics.git\ncd mcp-victoriametrics\ngo build -o bin/mcp-victoriametrics ./cmd/mcp-victoriametrics/main.go\n\n# after that add bin/mcp-victoriametrics file to your PATH\n```\n\n### Binärdateien\n\nLaden Sie einfach die neueste Version von der Seite [„Releases](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases) “ herunter und fügen Sie sie in Ihren PATH ein.\n\n### Docker\n\nDemnächst verfügbar...\n\n## Konfiguration\n\nDer MCP-Server für VictoriaMetrics wird über Umgebungsvariablen konfiguriert:\n\n| Variable | Beschreibung | Erforderlich | Standard | Zulässige Werte |\n| -------------------------- | --------------------------------------------------- | ------------ | ---------------- | -------------------- |\n| `VM_INSTANCE_ENTRYPOINT` | URL zur VictoriaMetrics-Instanz | Ja | - | - |\n| `VM_INSTANCE_TYPE` | Typ der VictoriaMetrics-Instanz | Ja | - | `single` , `cluster` |\n| `VM_INSTANCE_BEARER_TOKEN` | Authentifizierungstoken für die VictoriaMetrics-API | NEIN | - | - |\n| `MCP_SERVER_MODE` | Serverbetriebsmodus | NEIN | `stdio` | `stdio` , `sse` |\n| `MCP_SSE_ADDR` | Adresse, auf der der SSE-Server lauschen soll | NEIN | `localhost:8080` | - |\n\n### Konfigurationsbeispiele\n\n```bash\n# For a single-node instance\nexport VM_INSTANCE_ENTRYPOINT=\"http://localhost:8428\"\nexport VM_INSTANCE_TYPE=\"single\"\nexport VM_INSTANCE_BEARER_TOKEN=\"your-token\"\n\n# For a cluster\nexport VM_INSTANCE_ENTRYPOINT=\"https://play.victoriametrics.com\"\nexport VM_INSTANCE_TYPE=\"cluster\"\n\n# Server mode\nexport MCP_SERVER_MODE=\"sse\"\nexport MCP_SSE_ADDR=\"0.0.0.0:8080\"\n```\n\n## Einrichtung in Clients\n\n### Cursor\n\nGehen Sie zu: `Settings` -> `Cursor Settings` -> `MCP` -> `Add new global MCP server` und fügen Sie die folgende Konfiguration in Ihre Cursordatei `~/.cursor/mcp.json` ein:\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\nWeitere Informationen finden Sie in [den Cursor-MCP-Dokumenten](https://docs.cursor.com/context/model-context-protocol) .\n\n### Claude Desktop\n\nFügen Sie dies zu Ihrer Claude Desktop-Datei `claude_desktop_config.json` hinzu (Sie finden es, wenn Sie `Settings` -> `Developer` -> `Edit config` bearbeiten öffnen):\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\nWeitere Informationen finden Sie in [den Claude Desktop MCP-Dokumenten](https://modelcontextprotocol.io/quickstart/user) .\n\n### Claude Code\n\nFühren Sie den folgenden Befehl aus:\n\n```sh\nclaude mcp add victoriametrics -- /path/to/mcp-victoriametrics \\\n -e VM_INSTANCE_ENTRYPOINT= \\\n -e VM_INSTANCE_TYPE=\n -e VM_INSTANCE_BEARER_TOKEN=\n```\n\nWeitere Informationen finden Sie in [den MCP-Dokumenten von Claude Code](https://docs.anthropic.com/en/docs/agents-and-tools/claude-code/tutorials#set-up-model-context-protocol-mcp) .\n\n### Visual Studio Code\n\nFügen Sie dies zu Ihrer VS Code MCP-Konfigurationsdatei hinzu:\n\n```json\n{\n \"servers\": {\n \"victoriametrics\": {\n \"type\": \"stdio\",\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\nWeitere Informationen finden Sie in [den VS Code MCP-Dokumenten](https://code.visualstudio.com/docs/copilot/chat/mcp-servers) .\n\n### Zed\n\nFügen Sie Ihrer Zed-Konfigurationsdatei Folgendes hinzu:\n\n```json\n \"context_servers\": {\n \"victoriametrics\": {\n \"command\": {\n \"path\": \"/path/to/mcp-victoriametrics\",\n \"args\": [],\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n },\n \"settings\": {}\n }\n }\n```\n\nWeitere Informationen finden Sie in [den Zed MCP-Dokumenten](https://zed.dev/docs/ai/mcp) .\n\n### JetBrains IDEs\n\n* Öffnen Sie `Settings` -> `Tools` -> `AI Assistant` -> `Model Context Protocol (MCP)` .\n* Klicken Sie `Add (+)`\n* `As JSON` auswählen\n* Tragen Sie in das Eingabefeld Folgendes ein:\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n### Windsurf\n\nFügen Sie Ihrer Windsurf MCP-Konfigurationsdatei Folgendes hinzu.\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\nWeitere Informationen finden Sie in [den Windsurf MCP-Dokumenten](https://docs.windsurf.com/windsurf/mcp) .\n\n### Amazon Bedrock\n\nDemnächst verfügbar...\n\n### Verwenden von Docker anstelle von Binärdateien\n\nDemnächst verfügbar...\n\n## Verwendung\n\nNachdem Sie den MCP-Server [installiert](#installation) und [konfiguriert](#setup-in-clients) haben, können Sie ihn mit Ihrem bevorzugten MCP-Client verwenden.\n\nSie können den Dialog mit dem KI-Assistenten mit folgendem Satz starten:\n\n```\nUse MCP VictoriaMetrics in the following answers\n```\n\nDies ist jedoch nicht erforderlich. Sie können einfach Fragen stellen, und der Assistent verwendet automatisch die Tools und die Dokumentation, um Ihnen die besten Antworten zu liefern. Werfen Sie einen Blick in [die Dialogbeispiele,](#dialog-example) um besser zu verstehen, was Sie damit tun können.\n\n### Werkzeugsatz\n\nMCP VictoriaMetrics bietet zahlreiche Tools für die Interaktion mit Ihrer VictoriaMetrics-Instanz.\n\nHier ist eine Liste der verfügbaren Tools:\n\n| Werkzeug | Beschreibung |\n| ---------------------------- | --------------------------------------------------------------------------- |\n| `query` | Führen Sie sofortige PromQL/MetricsQL-Abfragen aus |\n| `query_range` | Führen Sie über einen bestimmten Zeitraum PromQL/MetricsQL-Abfragen aus |\n| `metrics` | Auflisten der verfügbaren Metriken |\n| `labels` | Auflisten der verfügbaren Etikettennamen |\n| `label_values` | Listenwerte für ein bestimmtes Label |\n| `series` | Verfügbare Zeitreihen auflisten |\n| `export` | Exportieren Sie Rohdaten der Zeitreihen in JSON oder CSV |\n| `rules` | Alarmierungs- und Aufzeichnungsregeln anzeigen |\n| `alerts` | Aktuelle Warnungen anzeigen (ausgelöst und ausstehend) |\n| `flags` | Nicht standardmäßige Flags der VictoriaMetrics-Instanz anzeigen |\n| `metric_statistics` | Statistiken zur Metriknutzung (in Abfragen) abrufen |\n| `active_queries` | Anzeigen aktuell ausgeführter Abfragen |\n| `top_queries` | Häufigste oder langsamste Abfragen anzeigen |\n| `tsdb_status` | Anzeigen der TSDB-Kardinalitätsstatistik |\n| `tenants` | Auflisten der verfügbaren Mandanten in der Multi-Tenant-Cluster-Einrichtung |\n| `documentation` | Suche in der eingebetteten VictoriaMetrics-Dokumentation |\n| `metric_relabel_debug` | Debuggen von Prometheus-kompatiblen Umbenennungsregeln |\n| `downsampling_filters_debug` | Debuggen der Downsampling-Konfiguration |\n| `retention_filters_debug` | Konfiguration der Debug-Aufbewahrungsfilter |\n\n### Eingabeaufforderungen\n\nDer Server enthält vordefinierte Eingabeaufforderungen für allgemeine Aufgaben.\n\nDies sind im Moment nur Beispiele, die Prompt-Bibliothek wird in Zukunft erweitert:\n\n| Prompt | Beschreibung |\n| ------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------- |\n| `unused_metrics` | Finden Sie Metriken, die nie abgefragt werden, und erstellen Sie Relabel-Konfigurationen, um deren Aufnahme zu stoppen |\n| `documentation` | Durchsuchen Sie die VictoriaMetrics-Dokumentation nach bestimmten Themen |\n| `rarely_used_metrics_with_high_cardinality` | Identifizieren Sie Metriken mit hoher Kardinalität, die selten abgefragt werden |\n\n## Dialogbeispiel\n\nDies ist nur ein Beispiel für einen Dialog mit dem Assistenten unter Verwendung des MCP-Servers, das zum Verständnis der Funktionsweise beitragen kann (es handelt sich jedoch um einen echten Dialog unter Verwendung des Victoria Metrics MCP-Servers mit Claude Desktop und Konfiguration als [öffentliche Playground-Instanz](https://play.victoriametrics.com) ):\n\n* > **Sie** : Welche Mandanten sind in meiner VM verfügbar?\n* > **Assistent** *In Ihrer VictoriaMetrics-Instanz sind die folgenden Mandanten verfügbar `0:0` (Standardmandant), `1:0` , `10:0` , `100:0`*\n* > **Sie** : Verwenden Sie ab sofort nur noch den Standardmandanten\n* > **Assistent** : *Ich verstehe. Von nun an verwende ich für alle Vorgänge mit Ihrer VictoriaMetrics-Instanz nur noch den Standardmandanten (0:0).*\n* > **Sie** : Welche Kennzahlen haben wir?\n* > **Assistent** : *In Ihrer VictoriaMetrics-Instanz sind zahlreiche Metriken verfügbar. Die Liste enthält Metriken aus verschiedenen Systemen und Komponenten, wie zum Beispiel:*\n >\n > * *Systemmetriken ( `node_` , `machine_` , usw.),*\n > * *Kubernetes-Metriken ( `kube_` , `container_` , `apiserver_` usw.),*\n > * *VictoriaMetrics-spezifische Metriken ( `vm_` , `vmagent_` , `vmalert_` usw.),*\n > * *Go-Laufzeitmetriken ( `go_` ),*\n > * *Prometheus-bezogene Metriken ( `scrape_` , `promhttp_` ),*\n > * *Verschiedene Anwendungsmetriken ( `argocd_` , `alertmanager_` ).*\n >\n > *Dies ist eine umfassende Liste mit über 2.000 Metriken, die für Abfragen zur Verfügung stehen. Gibt es einen bestimmten Metriktyp, den Sie genauer untersuchen möchten?*\n* > **Sie** : Okay, arbeiten wir ab jetzt mit Kubernetes-Metriken. Bitte zeigen Sie mir, welche dieser Metriken nie abgefragt wurden.\n* > **Assistent** : *Anhand der Ergebnisse kann ich erkennen, dass es zwei Kubernetes-Metriken gibt, die noch nie abgefragt wurden (Anzahl der Abfragen ist 0):*\n >\n > * *`kube_persistentvolumeclaim_deletion_timestamp` – Diese Metrik liefert Informationen darüber, wann PersistentVolumeClaims zum Löschen markiert wurden.*\n > * *`kube_pod_init_container_status_last_terminated_reason` – Diese Metrik zeigt den Grund, warum Init-Container in Pods zuletzt beendet wurden.*\n >\n > *Diese Messwerte können für Überwachungs- und Fehlerbehebungszwecke hilfreich sein:*\n >\n > * *Die Metrik des PVC-Löschzeitstempels kann dabei helfen, Löschmuster zu verfolgen oder feststeckende PVCs zu identifizieren.*\n > * *Die Metrik für den Grund der Beendigung des Init-Containers kann dabei helfen, häufige Fehlermuster in Ihren Initialisierungscontainern zu identifizieren.*\n >\n > *Möchten Sie, dass ich Ihnen mehr darüber erkläre, was diese Metriken verfolgen, oder einige nützliche Abfragen vorschlage, um mit der Überwachung zu beginnen?*\n* > **Sie** : Ja, bitte. Können Sie außerdem Regeln zur Neubezeichnung erstellen, um die Aufnahme dieser Messwerte zu unterbinden?\n* > **Assistent** : *Ich erstelle eine Relabel-Konfiguration, um diese nie abgefragten Metriken von der Erfassung auszuschließen. Dies reduziert den Speicher- und Verarbeitungsaufwand. Hier ist die Relabel-Konfiguration, die Sie Ihrer VictoriaMetrics-Scraping-Konfiguration hinzufügen können:*\n >\n > ```yaml\n > relabel_configs:\n > - action: drop\n > source_labels: [__name__]\n > regex: kube_persistentvolumeclaim_deletion_timestamp|kube_pod_init_container_status_last_terminated_reason\n > ```\n >\n > *Diese Konfiguration wird:*\n >\n > 1. *Schauen Sie sich den Namen jeder Metrik an (das Label `__name__` ).*\n > 2. *Überprüfen Sie, ob es einer der angegebenen Metriken entspricht*\n > 3. *Löschen Sie die übereinstimmenden Metriken und verhindern Sie so deren Erfassung und Speicherung.*\n >\n > *Sie können dies Ihrer VictoriaMetrics-Konfigurationsdatei oder den entsprechenden Scrape-Job-Konfigurationen in Ihrer Prometheus-Konfiguration hinzufügen, wenn Sie Prometheus als Scraper für VictoriaMetrics verwenden.* *Wenn Sie diese Relabel-Konfiguration vor der Implementierung testen möchten, können Sie das Debug-Tool für Metrik-Relabels verwenden, um die Auswirkungen auf Ihre Metriken zu prüfen.*\n* > **Sie** : Ja, bitte debuggen Sie diese Relabel-Konfiguration in Echtzeitreihen, um zu sehen, wie sie sich auf meine Metriken auswirken würde\n* > **Assistent** : *Perfekt! Die Neubeschriftungskonfiguration funktioniert wie erwartet:*\n >\n > * *Die ersten beiden Testfälle zeigen, dass sowohl die Metriken `kube_persistentvolumeclaim_deletion_timestamp` als auch `kube_pod_init_container_status_last_terminated_reason` korrekt gelöscht werden (die resultierenden Bezeichnungen sind leer `{}` ).*\n > * *Der dritte Testfall bestätigt, dass andere Metriken wie `kube_pod_status_phase` erhalten bleiben und von unserer Drop-Regel nicht beeinflusst werden.*\n >\n > *Sie können diese Relabel-Konfiguration problemlos zu Ihrer VictoriaMetrics- oder Prometheus-Konfiguration hinzufügen, um die Erfassung dieser ungenutzten Metriken zu beenden. Dies spart Speicherplatz und reduziert den Verarbeitungsaufwand, ohne Ihre aktuellen Überwachungsfunktionen zu beeinträchtigen.*\n* > **Sie** : Aber ich verwende die vmagent-Ressource des VictoriaMetrics-Operators. Könnten Sie mir bitte einen Teil der Spezifikation für meinen Agenten mit dieser Relabel-Konfiguration geben?\n* > **Assistent** : *Sie haben Recht. Da Sie den VictoriaMetrics-Operator und die benutzerdefinierte VMAgent-Ressource verwenden, muss die Konfiguration im entsprechenden Format für diese CR vorliegen. So fügen Sie die Relabel-Konfiguration zu Ihrer VMAgent-Spezifikation hinzu:*\n >\n > ```yaml\n > apiVersion: operator.victoriametrics.com/v1beta1\n > kind: VMAgent\n > metadata:\n > name: your-vmagent-name\n > spec:\n > # ... your existing configuration ...\n > \n > # Add this section to configure global metric relabeling\n > inlineRelabelConfig:\n > - action: drop\n > sourceLabels: [\"__name__\"]\n > regex: kube_persistentvolumeclaim_deletion_timestamp|kube_pod_init_container_status_last_terminated_reason\n > \n > # ... rest of your configuration ...\n > ```\n >\n > *Wichtige Punkte:*\n >\n > * *In der VMAgent-Spezifikation verwenden Sie `inlineRelabelConfig` für Inline-Relabeling-Regeln*\n > * *Das Feld heißt `sourceLabels` (nicht `source_labels` wie in der Prometheus-Konfiguration).*\n > * *Dies ist eine globale Neubezeichnung, die auf alle Metriken angewendet wird, bevor sie gespeichert werden*\n* > **Sie** : Ich habe beschlossen, diese Messwerte nicht zu entfernen, sondern lediglich die Abtastrate durch Downsampling zu senken. Welche Downsampling-Parameter habe ich jetzt?\n* > **Assistent** : *Ich sehe, dass Sie derzeit die folgende Downsampling-Konfiguration haben:*\n >\n > ```\n > -downsampling.period=\"7d:1m,30d:5m,180d:30m\"\n > ```\n >\n > *Das heisst:*\n >\n > * *Daten, die neuer als 7 Tage sind, werden in ihrer ursprünglichen Auflösung gespeichert*\n > * *Daten, die zwischen 7 und 30 Tage alt sind, werden auf eine Auflösung von 1 Minute heruntergerechnet*\n > * *Daten, die 30–180 Tage alt sind, werden auf eine Auflösung von 5 Minuten heruntergerechnet.*\n > * *Daten, die älter als 180 Tage sind, werden auf eine Auflösung von 30 Minuten heruntergerechnet.*\n >\n > *Wenn Sie nun spezifische Downsampling-Regeln für diese selten abgefragten Metriken hinzufügen möchten, können Sie mit dem Debug-Tool für Downsampling-Filter testen, wie die neuen Regeln funktionieren. Testen wir eine Konfiguration, bei der wir aggressiveres Downsampling auf diese spezifischen Metriken anwenden:*\n >\n > *...Serien abrufen und Downsampling-Filter-Debugger verwenden...*\n >\n > *Super! Unser Test funktioniert jetzt einwandfrei. Basierend auf der aktuellen Konfiguration und unseren Tests finden Sie hier eine empfohlene Downsampling-Konfiguration für Ihren VMAgent CR, die ein aggressiveres Downsampling auf die selten abgefragten Metriken anwendet:*\n >\n > ```yaml\n > apiVersion: operator.victoriametrics.com/v1beta1\n > kind: VMCluster\n > metadata:\n > name: monitoring-vm-cluster\n > spec:\n > # ... your existing configuration ...\n > \n > # Update the vmstorage section with the new downsampling parameters\n > vmstorage:\n > extraArgs:\n > downsampling.period: '{__name__=\\\"kube_persistentvolumeclaim_deletion_timestamp\\\"}:1d:10m,{__name__=\\\"kube_pod_init_container_status_last_terminated_reason\\\"}:1d:10m,7d:1m,30d:5m,180d:30m'\n > \n > # ... rest of your configuration ...\n > ```\n >\n > *Diese Konfiguration:*\n >\n > *- Wendet für die selten abgefragten Metriken nach nur einem Tag ein Downsampling mit einer Auflösung von 10 Minuten an* *- Behält Ihre bestehende Downsampling-Richtlinie für alle anderen Metriken bei*\n >\n > *Die Reihenfolge der Downsampling-Regeln ist wichtig – spezifischere Regeln sollten zuerst aufgeführt werden.*\n\nWährend dieses Dialogs verwendete der Assistent die folgenden Tools:\n\n* `tenants` , um die Liste der verfügbaren Mieter zu erhalten\n* `documentation` , um Informationen zu Funktionalität und Datenformaten zu erhalten\n* `metrics` , um die Liste der verfügbaren Metriken abzurufen\n* `metrics_staistics` um Informationen zur Metriknutzung zu erhalten\n* `series` , um die Zeitreihen zum Debuggen zu erhalten\n* `metric_relabel_debug` zum Debuggen von Relabeling-Regeln\n* `flags` zum Abrufen von Informationen zu Instanzparametern\n* `downsampling_filters_debug` zum Debuggen der Downsampling-Konfiguration\n\nSie können aber auch beliebige andere Tools verwenden und diese auf Ihre eigene Weise kombinieren.\n\n## Haftungsausschluss\n\nKI-Dienste und -Agenten sowie MCP-Server wie dieser können die Genauigkeit, Vollständigkeit und Zuverlässigkeit der Ergebnisse nicht garantieren. Sie sollten die mit KI erzielten Ergebnisse überprüfen. Die Qualität des MCP-Servers und seiner Antworten hängt stark von den Fähigkeiten Ihres Clients und der Qualität des verwendeten Modells ab.\n\n## Beitragen\n\nBeiträge zum MCP VictoriaMetrics-Projekt sind willkommen! Senden Sie uns gerne Probleme, Funktionswünsche oder Pull Requests.","German","tag","de-DE",{"_27":38,"_31":40,"_35":41},{"_29":30,"_31":30,"_32":39},"# VictoriaMetrics MCP Server\n\nThe implementation of [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) server for [VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/).\n\nThis provides access to your VictoriaMetrics instance and seamless integration with [VictoriaMetrics APIs](https://docs.victoriametrics.com/victoriametrics/url-examples/) and [documentation](https://docs.victoriametrics.com/).\nIt can give you a comprehensive interface for monitoring, observability, and debugging tasks related to your VictoriaMetrics instances, enable advanced automation and interaction capabilities for engineers and tools.\n\n## Features\n\nThis MCP server allows you to use almost all read-only APIs of VictoriaMetrics, i.e. all functions available in [VMUI](https://docs.victoriametrics.com/#vmui):\n\n- Querying metrics and exploring data (even drawing graphs if your client supports it)\n- Listing and exporting available metrics, labels, labels values and entire series\n- Analyzing your alerting and recording rules and alerts\n- Showing parameters of your VictoriaMetrics instance\n- Exploring cardinality of your data and metrics usage statistics\n- Analyzing your queries\n- Debugging your relabeling rules, downsampling and retention policy configurations \n \nIn addition, the MCP server contains embedded up-to-date documentation and is able to search it without online access.\n\nMore details about the exact available tools and prompts can be found in the [Usage](#usage) section.\n\nYou can combine functionality of tools, docs search in your prompts and invent great usage scenarios for your VictoriaMetrics instance.\nJust check the [Dialog example](#dialog-example) section to see how it can work.\nAnd please note the fact that the quality of the MCP Server and its responses depends very much on the capabilities of your client and the quality of the model you are using.\n\nYou can also combine the MCP server with other observability or doc search related MCP Servers and get even more powerful results.\n\n## Requirements\n\n- [VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/) instance ([single-node](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/) or [cluster](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/))\n- Go 1.24 or higher (if you want to build from source)\n\n## Installation\n\n### Go\n\n```bash\ngo install github.com/VictoriaMetrics-Community/mcp-victoriametrics/cmd/mcp-victoriametrics@latest\n```\n\n### Source Code\n\n```bash\ngit clone https://github.com/VictoriaMetrics-Community/mcp-victoriametrics.git\ncd mcp-victoriametrics\ngo build -o bin/mcp-victoriametrics ./cmd/mcp-victoriametrics/main.go\n\n# after that add bin/mcp-victoriametrics file to your PATH\n```\n\n### Binaries\n\nJust download the latest release from [Releases](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases) page and put it to your PATH.\n\n### Docker\n\nComing soon...\n\n\n\n## Configuration\n\nMCP Server for VictoriaMetrics is configured via environment variables:\n\n| Variable | Description | Required | Default | Allowed values |\n|----------|-------------|----------|---------|---------|\n| `VM_INSTANCE_ENTRYPOINT` | URL to VictoriaMetrics instance | Yes | - | - |\n| `VM_INSTANCE_TYPE` | Type of VictoriaMetrics instance | Yes | - | `single`, `cluster` |\n| `VM_INSTANCE_BEARER_TOKEN` | Authentication token for VictoriaMetrics API | No | - | - |\n| `MCP_SERVER_MODE` | Server operation mode | No | `stdio` | `stdio`, `sse` |\n| `MCP_SSE_ADDR` | Address for SSE server to listen on | No | `localhost:8080` | - |\n\n### Сonfiguration examples\n\n```bash\n# For a single-node instance\nexport VM_INSTANCE_ENTRYPOINT=\"http://localhost:8428\"\nexport VM_INSTANCE_TYPE=\"single\"\nexport VM_INSTANCE_BEARER_TOKEN=\"your-token\"\n\n# For a cluster\nexport VM_INSTANCE_ENTRYPOINT=\"https://play.victoriametrics.com\"\nexport VM_INSTANCE_TYPE=\"cluster\"\n\n# Server mode\nexport MCP_SERVER_MODE=\"sse\"\nexport MCP_SSE_ADDR=\"0.0.0.0:8080\"\n```\n\n## Setup in clients\n\n### Cursor\n\nGo to: `Settings` -> `Cursor Settings` -> `MCP` -> `Add new global MCP server` and paste the following configuration into your Cursor `~/.cursor/mcp.json` file:\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\nSee [Cursor MCP docs](https://docs.cursor.com/context/model-context-protocol) for more info.\n\n### Claude Desktop\n\nAdd this to your Claude Desktop `claude_desktop_config.json` file (you can find it if open `Settings` -> `Developer` -> `Edit config`):\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\nSee [Claude Desktop MCP docs](https://modelcontextprotocol.io/quickstart/user) for more info.\n\n### Claude Code\n\nRun the command:\n\n```sh\nclaude mcp add victoriametrics -- /path/to/mcp-victoriametrics \\\n -e VM_INSTANCE_ENTRYPOINT= \\\n -e VM_INSTANCE_TYPE=\n -e VM_INSTANCE_BEARER_TOKEN=\n```\n\nSee [Claude Code MCP docs](https://docs.anthropic.com/en/docs/agents-and-tools/claude-code/tutorials#set-up-model-context-protocol-mcp) for more info.\n\n### Visual Studio Code\n\nAdd this to your VS Code MCP config file:\n\n```json\n{\n \"servers\": {\n \"victoriametrics\": {\n \"type\": \"stdio\",\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\nSee [VS Code MCP docs](https://code.visualstudio.com/docs/copilot/chat/mcp-servers) for more info.\n\n### Zed\n\nAdd the following to your Zed config file:\n\n```json\n \"context_servers\": {\n \"victoriametrics\": {\n \"command\": {\n \"path\": \"/path/to/mcp-victoriametrics\",\n \"args\": [],\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n },\n \"settings\": {}\n }\n }\n```\n\nSee [Zed MCP docs](https://zed.dev/docs/ai/mcp) for more info.\n\n### JetBrains IDEs\n\n- Open `Settings` -> `Tools` -> `AI Assistant` -> `Model Context Protocol (MCP)`.\n- Click `Add (+)`\n- Select `As JSON`\n- Put the following to the input field:\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n### Windsurf\n\nAdd the following to your Windsurf MCP config file.\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\nSee [Windsurf MCP docs](https://docs.windsurf.com/windsurf/mcp) for more info.\n\n### Amazon Bedrock\n\nComing soon....\n\n### Using Docker instead of binary\n\nComing soon...\n\n## Usage\n\nAfter [installing](#installation) and [configuring](#setup-in-clients) the MCP server, you can start using it with your favorite MCP client.\n\nYou can start dialog with AI assistant from the phrase:\n\n```\nUse MCP VictoriaMetrics in the following answers\n```\n\nBut it's not required, you can just start asking questions and the assistant will automatically use the tools and documentation to provide you with the best answers.\nJust take a look into [Dialog example](#dialog-example) section for better understanding what you can do with it.\n\n### Toolset\n\nMCP VictoriaMetrics provides numerous tools for interacting with your VictoriaMetrics instance.\n\nHere's a list of available tools:\n\n| Tool | Description |\n|------|-----------------------------------------------------------|\n| `query` | Execute instant PromQL/MetricsQL queries |\n| `query_range` | Execute range PromQL/MetricsQL queries over a time period |\n| `metrics` | List available metrics |\n| `labels` | List available label names |\n| `label_values` | List values for a specific label |\n| `series` | List available time series |\n| `export` | Export raw time series data to JSON or CSV |\n| `rules` | View alerting and recording rules |\n| `alerts` | View current alerts (firing and pending) |\n| `flags` | View non-default flags of the VictoriaMetrics instance |\n| `metric_statistics` | Get metrics usage (in queries) statistics |\n| `active_queries` | View currently executing queries |\n| `top_queries` | View most frequent or slowest queries |\n| `tsdb_status` | View TSDB cardinality statistics |\n| `tenants` | List available tenants in multi-tenant cluster setup |\n| `documentation` | Search in embedded VictoriaMetrics documentation |\n| `metric_relabel_debug` | Debug Prometheus-compatible relabeling rules |\n| `downsampling_filters_debug` | Debug downsampling configuration |\n| `retention_filters_debug` | Debug retention filters configuration |\n\n### Prompts\n\nThe server includes pre-defined prompts for common tasks.\n\nThese are just examples at the moment, the prompt library will be added to in the future:\n\n| Prompt | Description |\n|--------|-------------|\n| `unused_metrics` | Find metrics that are never queried and create relabel configs to stop ingesting them |\n| `documentation` | Search VictoriaMetrics documentation for specific topics |\n| `rarely_used_metrics_with_high_cardinality` | Identify high-cardinality metrics that are rarely queried |\n\n## Dialog example\n\nThis is just example of dialog with Assistant using MCP Server that can help to understand how it can work (but's it's real dialog with using the Victoria Metrics MCP server with Claude Desktop and configured to [public playground instance](https://play.victoriametrics.com)):\n\n- > **You**: What tenants are available in my VM?\n- > **Assistant** *In your VictoriaMetrics instance, the following tenants are available `0:0` (default tenant), `1:0`, `10:0`, `100:0`*\n- > **You**: From now on, please use only default tenant\n- > **Assistant**: *I understand. From now on, I'll use only the default tenant (0:0) for all operations with your VictoriaMetrics instance.*\n- > **You**: Which metrics do we have?\n- > **Assistant**: *There are many metrics available in your VictoriaMetrics instance. The list includes metrics from various systems and components, such as:*\n >\n > - *System metrics (`node_`, `machine_`, etc.),*\n > - *Kubernetes metrics (`kube_`, `container_`, `apiserver_`, etc.),*\n > - *VictoriaMetrics-specific metrics (`vm_`, `vmagent_`, `vmalert_`, etc.),*\n > - *Go runtime metrics (`go_`),*\n > - *Prometheus-related metrics (`scrape_`, `promhttp_`),*\n > - *Various application metrics (`argocd_`, `alertmanager_`).*\n >\n > *This is a comprehensive list of over 2000 metrics that are available for querying. Is there a specific type of metric you're interested in exploring further?*\n- > **You**: Ok, let's work with kubernetes metrics from now on. Please show me which of these metrics never queried?\n- > **Assistant**: *Based on the results, I can see that there are two Kubernetes metrics that have never been queried (query count is 0):*\n >\n > - *`kube_persistentvolumeclaim_deletion_timestamp` - This metric provides information about when PersistentVolumeClaims were marked for deletion.*\n > - *`kube_pod_init_container_status_last_terminated_reason` - This metric shows the reason why init containers in pods were last terminated.*\n >\n > *These metrics might be valuable for monitoring and troubleshooting purposes:*\n >\n > - *The PVC deletion timestamp metric could help track deletion patterns or identify stuck PVCs.*\n > - *The init container termination reason metric could help identify common failure patterns in your initialization containers.*\n >\n > *Would you like me to explain more about what these metrics track or suggest some useful queries to start monitoring them?*\n- > **You**: Yes, please. And also, can you create relabeling rules to stop ingesting these metrics?\n- > **Assistant**: *I'll create a relabel configuration to exclude these never-queried metrics from collection. This will help reduce storage and processing overhead. Here's the relabel configuration you can add to your VictoriaMetrics scrape configuration:*\n >\n > ```yaml\n > relabel_configs:\n > - action: drop\n > source_labels: [__name__]\n > regex: kube_persistentvolumeclaim_deletion_timestamp|kube_pod_init_container_status_last_terminated_reason\n > ```\n >\n > *This configuration will:*\n > 1. *Look at each metric's name (the `__name__` label)*\n > 2. *Check if it matches either of the specified metrics*\n > 3. *Drop the metrics that match, preventing them from being collected and stored*\n >\n > *You can add this to your VictoriaMetrics configuration file or to the relevant scrape job configurations in your Prometheus configuration if you're using Prometheus as a scraper for VictoriaMetrics.*\n > *If you want to test this relabel configuration before implementing it, you can use the metric relabel debug tool to see how it would affect your metrics.*\n- > **You**: Yes, please debug this relabel config on real time series to see how it would affect my metrics\n- > **Assistant**: *Perfect! The relabel config works as expected:*\n >\n > - *The first two test cases show that both `kube_persistentvolumeclaim_deletion_timestamp` and `kube_pod_init_container_status_last_terminated_reason` metrics are correctly dropped (resulting labels are empty `{}`).*\n > - *The third test case confirms that other metrics like `kube_pod_status_phase` are preserved and not affected by our drop rule.*\n >\n > *You can confidently add this relabel configuration to your VictoriaMetrics or Prometheus configuration to stop collecting these unused metrics. This will save storage space and reduce processing overhead without affecting your current monitoring capabilities.*\n- > **You**: But i'm using vmagent resource of VictoriaMetrics operator, could you please give me part of specification for my agent with this relabel config?\n- > **Assistant**: *You're right, since you're using the VictoriaMetrics Operator and the VMAgent custom resource, the configuration needs to be in the appropriate format for that CR. Here's how you can add the relabel configuration to your VMAgent specification:*\n >\n > ```yaml\n > apiVersion: operator.victoriametrics.com/v1beta1\n > kind: VMAgent\n > metadata:\n > name: your-vmagent-name\n > spec:\n > # ... your existing configuration ...\n > \n > # Add this section to configure global metric relabeling\n > inlineRelabelConfig:\n > - action: drop\n > sourceLabels: [\"__name__\"]\n > regex: kube_persistentvolumeclaim_deletion_timestamp|kube_pod_init_container_status_last_terminated_reason\n > \n > # ... rest of your configuration ...\n > ```\n >\n > *Key points:*\n >\n > - *In the VMAgent spec, you use `inlineRelabelConfig` for inline relabeling rules*\n > - *The field is called `sourceLabels` (not `source_labels` as in Prometheus config)*\n > - *This is a global relabeling that will be applied to all metrics before they're stored*\n- > **You**: I decided not to remove those metrics, but just lower the sample rate for them using Downsampling. What downsampling parameters i have now?\n- > **Assistant**: *I can see that you currently have the following downsampling configuration:*\n >\n > ```\n > -downsampling.period=\"7d:1m,30d:5m,180d:30m\"\n > ```\n >\n > *This means:*\n >\n > - *Data newer than 7 days is kept at its original resolution*\n > - *Data between 7-30 days old is downsampled to 1-minute resolution*\n > - *Data between 30-180 days old is downsampled to 5-minute resolution*\n > - *Data older than 180 days is downsampled to 30-minute resolution*\n >\n > *Now, if you want to add specific downsampling rules for those rarely-queried metrics, you can use the downsampling filters debug tool to test how new rules would work. Let's test a configuration where we apply more aggressive downsampling to those specific metrics:*\n >\n > *...getting series and using downsampling filter debugger...*\n >\n > *Great! Now our test is working correctly. Based on the current configuration and our testing, here's a suggested downsampling configuration for your VMAgent CR that would apply more aggressive downsampling to those rarely-queried metrics:*\n >\n > ```yaml\n > apiVersion: operator.victoriametrics.com/v1beta1\n > kind: VMCluster\n > metadata:\n > name: monitoring-vm-cluster\n > spec:\n > # ... your existing configuration ...\n > \n > # Update the vmstorage section with the new downsampling parameters\n > vmstorage:\n > extraArgs:\n > downsampling.period: '{__name__=\\\"kube_persistentvolumeclaim_deletion_timestamp\\\"}:1d:10m,{__name__=\\\"kube_pod_init_container_status_last_terminated_reason\\\"}:1d:10m,7d:1m,30d:5m,180d:30m'\n > \n > # ... rest of your configuration ...\n > ``` \n >\n > *This configuration:*\n >\n > *- Applies 10-minute resolution downsampling for the rarely-queried metrics after just 1 day*\n > *- Maintains your existing downsampling policy for all other metrics*\n >\n > *The order of the downsampling rules is important - more specific rules should be listed first.*\n\nDuring this dialog, the assistant was using the following tools:\n\n- `tenants` to get the list of available tenants\n- `documentation` to get information about functionality and data formats\n- `metrics` to get the list of available metrics\n- `metrics_staistics` to get the information about metrics usage\n- `series` to get the time series for debugging\n- `metric_relabel_debug` to debug relabeling rules\n- `flags` to get the information about instance parameters\n- `downsampling_filters_debug` to debug downsampling configuration\n\nBut you can use any other tools and combine them in your own way.\n\n## Disclaimer\n\nAI services and agents along with MCP servers like this cannot guarantee the accuracy, completeness and reliability of results.\nYou should double check the results obtained with AI.\nThe quality of the MCP Server and its responses depends very much on the capabilities of your client and the quality of the model you are using.\n\n## Contributing\n\nContributions to the MCP VictoriaMetrics project are welcome! Please feel free to submit issues, feature requests, or pull requests.\n","English","en-US",{"_27":43,"_31":45,"_35":46},{"_29":30,"_31":30,"_32":44},"# Servidor MCP de VictoriaMetrics\n\nLa implementación del servidor [de Protocolo de Contexto de Modelo (MCP)](https://modelcontextprotocol.io/) para [VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/) .\n\nEsto proporciona acceso a su instancia de VictoriaMetrics y una integración fluida con las API y [la documentación](https://docs.victoriametrics.com/) [de VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/url-examples/) . Le ofrece una interfaz completa para las tareas de monitoreo, observabilidad y depuración relacionadas con sus instancias de VictoriaMetrics, además de habilitar capacidades avanzadas de automatización e interacción para ingenieros y herramientas.\n\n## Características\n\nEste servidor MCP le permite utilizar casi todas las API de solo lectura de VictoriaMetrics, es decir, todas las funciones disponibles en [VMUI](https://docs.victoriametrics.com/#vmui) :\n\n* Consultar métricas y explorar datos (incluso dibujar gráficos si su cliente lo admite)\n* Listado y exportación de métricas, etiquetas, valores de etiquetas y series completas disponibles\n* Análisis de sus reglas y alertas de grabación y alertas\n* Mostrar parámetros de su instancia de VictoriaMetrics\n* Explorando la cardinalidad de sus datos y estadísticas de uso de métricas\n* Analizando sus consultas\n* Depuración de sus reglas de reetiquetado, reducción de muestreo y configuraciones de políticas de retención\n\nAdemás, el servidor MCP contiene documentación actualizada incorporada y puede buscarla sin acceso en línea.\n\nPuede encontrar más detalles sobre las herramientas y indicaciones exactas disponibles en la sección [Uso](#usage) .\n\nPuede combinar la funcionalidad de herramientas y la búsqueda de documentos en sus indicaciones e idear escenarios de uso excelentes para su instancia de VictoriaMetrics. Simplemente consulte la sección [de ejemplo de diálogo](#dialog-example) para ver cómo funciona. Tenga en cuenta que la calidad del servidor MCP y sus respuestas dependen en gran medida de las capacidades de su cliente y de la calidad del modelo que utilice.\n\nTambién puede combinar el servidor MCP con otros servidores MCP relacionados con la observación o la búsqueda de documentos y obtener resultados aún más potentes.\n\n## Requisitos\n\n* Instancia [de VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/) ( [nodo único](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/) o [clúster](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/) )\n* Vaya a 1.24 o superior (si desea compilar desde la fuente)\n\n## Instalación\n\n### Ir\n\n```bash\ngo install github.com/VictoriaMetrics-Community/mcp-victoriametrics/cmd/mcp-victoriametrics@latest\n```\n\n### Código fuente\n\n```bash\ngit clone https://github.com/VictoriaMetrics-Community/mcp-victoriametrics.git\ncd mcp-victoriametrics\ngo build -o bin/mcp-victoriametrics ./cmd/mcp-victoriametrics/main.go\n\n# after that add bin/mcp-victoriametrics file to your PATH\n```\n\n### Binarios\n\nSimplemente descargue la última versión desde la página de [versiones](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases) y colóquela en su RUTA.\n\n### Estibador\n\nMuy pronto...\n\n## Configuración\n\nEl servidor MCP para VictoriaMetrics se configura a través de variables de entorno:\n\n| Variable | Descripción | Requerido | Por defecto | Valores permitidos |\n| -------------------------- | ----------------------------------------------------- | --------- | ---------------- | -------------------- |\n| `VM_INSTANCE_ENTRYPOINT` | URL a la instancia de VictoriaMetrics | Sí | - | - |\n| `VM_INSTANCE_TYPE` | Tipo de instancia de VictoriaMetrics | Sí | - | `single` , `cluster` |\n| `VM_INSTANCE_BEARER_TOKEN` | Token de autenticación para la API de VictoriaMetrics | No | - | - |\n| `MCP_SERVER_MODE` | Modo de funcionamiento del servidor | No | `stdio` | `stdio` , `sse` |\n| `MCP_SSE_ADDR` | Dirección para que el servidor SSE escuche | No | `localhost:8080` | - |\n\n### Ejemplos de configuración\n\n```bash\n# For a single-node instance\nexport VM_INSTANCE_ENTRYPOINT=\"http://localhost:8428\"\nexport VM_INSTANCE_TYPE=\"single\"\nexport VM_INSTANCE_BEARER_TOKEN=\"your-token\"\n\n# For a cluster\nexport VM_INSTANCE_ENTRYPOINT=\"https://play.victoriametrics.com\"\nexport VM_INSTANCE_TYPE=\"cluster\"\n\n# Server mode\nexport MCP_SERVER_MODE=\"sse\"\nexport MCP_SSE_ADDR=\"0.0.0.0:8080\"\n```\n\n## Configuración en clientes\n\n### Cursor\n\nVaya a: `Settings` -> `Cursor Settings` -> `MCP` -> `Add new global MCP server` y pegue la siguiente configuración en su archivo Cursor `~/.cursor/mcp.json` :\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\nConsulte [la documentación de Cursor MCP](https://docs.cursor.com/context/model-context-protocol) para obtener más información.\n\n### Escritorio de Claude\n\nAgregue esto a su archivo `claude_desktop_config.json` de Claude Desktop (puede encontrarlo si abre `Settings` -> `Developer` -> `Edit config` ):\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\nConsulte [la documentación de Claude Desktop MCP](https://modelcontextprotocol.io/quickstart/user) para obtener más información.\n\n### Claude Code\n\nEjecute el comando:\n\n```sh\nclaude mcp add victoriametrics -- /path/to/mcp-victoriametrics \\\n -e VM_INSTANCE_ENTRYPOINT= \\\n -e VM_INSTANCE_TYPE=\n -e VM_INSTANCE_BEARER_TOKEN=\n```\n\nConsulte [la documentación de Claude Code MCP](https://docs.anthropic.com/en/docs/agents-and-tools/claude-code/tutorials#set-up-model-context-protocol-mcp) para obtener más información.\n\n### Código de Visual Studio\n\nAgregue esto a su archivo de configuración MCP de VS Code:\n\n```json\n{\n \"servers\": {\n \"victoriametrics\": {\n \"type\": \"stdio\",\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\nConsulte [la documentación de VS Code MCP](https://code.visualstudio.com/docs/copilot/chat/mcp-servers) para obtener más información.\n\n### Zed\n\nAgregue lo siguiente a su archivo de configuración de Zed:\n\n```json\n \"context_servers\": {\n \"victoriametrics\": {\n \"command\": {\n \"path\": \"/path/to/mcp-victoriametrics\",\n \"args\": [],\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n },\n \"settings\": {}\n }\n }\n```\n\nConsulte [la documentación de Zed MCP](https://zed.dev/docs/ai/mcp) para obtener más información.\n\n### IDE de JetBrains\n\n* Abra `Settings` -> `Tools` -> `AI Assistant` -> `Model Context Protocol (MCP)` .\n* Haga clic `Add (+)`\n* Seleccionar `As JSON`\n* Coloque lo siguiente en el campo de entrada:\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n### Windsurf\n\nAgregue lo siguiente a su archivo de configuración Windsurf MCP.\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\nConsulte [la documentación de Windsurf MCP](https://docs.windsurf.com/windsurf/mcp) para obtener más información.\n\n### Amazon Bedrock\n\nMuy pronto....\n\n### Usando Docker en lugar de binario\n\nMuy pronto...\n\n## Uso\n\nDespués de [instalar](#installation) y [configurar](#setup-in-clients) el servidor MCP, puede comenzar a usarlo con su cliente MCP favorito.\n\nPuedes iniciar un diálogo con el asistente de IA desde la frase:\n\n```\nUse MCP VictoriaMetrics in the following answers\n```\n\nPero no es obligatorio; puedes empezar a hacer preguntas y el asistente usará automáticamente las herramientas y la documentación para ofrecerte las mejores respuestas. Consulta la sección [de ejemplos de Dialog](#dialog-example) para comprender mejor sus funciones.\n\n### Conjunto de herramientas\n\nMCP VictoriaMetrics proporciona numerosas herramientas para interactuar con su instancia de VictoriaMetrics.\n\nAquí hay una lista de herramientas disponibles:\n\n| Herramienta | Descripción |\n| ---------------------------- | ------------------------------------------------------------------------------------ |\n| `query` | Ejecutar consultas instantáneas de PromQL/MetricsQL |\n| `query_range` | Ejecutar consultas PromQL/MetricsQL de rango durante un período de tiempo |\n| `metrics` | Lista de métricas disponibles |\n| `labels` | Lista de nombres de etiquetas disponibles |\n| `label_values` | Lista de valores para una etiqueta específica |\n| `series` | Lista de series temporales disponibles |\n| `export` | Exportar datos de series de tiempo sin procesar a JSON o CSV |\n| `rules` | Ver reglas de alerta y grabación |\n| `alerts` | Ver alertas actuales (activadas y pendientes) |\n| `flags` | Ver indicadores no predeterminados de la instancia VictoriaMetrics |\n| `metric_statistics` | Obtener estadísticas de uso de métricas (en consultas) |\n| `active_queries` | Ver las consultas que se están ejecutando actualmente |\n| `top_queries` | Ver las consultas más frecuentes o más lentas |\n| `tsdb_status` | Ver las estadísticas de cardinalidad de TSDB |\n| `tenants` | Enumerar los inquilinos disponibles en la configuración de un clúster multiinquilino |\n| `documentation` | Buscar en la documentación integrada de VictoriaMetrics |\n| `metric_relabel_debug` | Depurar reglas de reetiquetado compatibles con Prometheus |\n| `downsampling_filters_debug` | Configuración de depuración de disminución de muestreo |\n| `retention_filters_debug` | Configuración de filtros de retención de depuración |\n\n### Indicaciones\n\nEl servidor incluye indicaciones predefinidas para tareas comunes.\n\nEstos son solo ejemplos por el momento, la biblioteca de indicaciones se agregará en el futuro:\n\n| Inmediato | Descripción |\n| ------------------------------------------- | --------------------------------------------------------------------------------------------------------- |\n| `unused_metrics` | Encuentre métricas que nunca se consultan y cree configuraciones de reetiquetado para dejar de ingerirlas |\n| `documentation` | Busque documentación de VictoriaMetrics para temas específicos |\n| `rarely_used_metrics_with_high_cardinality` | Identificar métricas de alta cardinalidad que rara vez se consultan |\n\n## Ejemplo de diálogo\n\nEste es solo un ejemplo de diálogo con el Asistente que usa el servidor MCP que puede ayudar a comprender cómo puede funcionar (pero es un diálogo real con el uso del servidor MCP de Victoria Metrics con Claude Desktop y configurado para [una instancia de juegos pública](https://play.victoriametrics.com) ):\n\n* > **Tú** : ¿Qué inquilinos están disponibles en mi máquina virtual?\n* > **Asistente** *En su instancia de VictoriaMetrics, los siguientes inquilinos están disponibles `0:0` (inquilino predeterminado), `1:0` , `10:0` , `100:0`*\n* > **Tú** : De ahora en adelante, utiliza únicamente el inquilino predeterminado\n* > **Asistente** : *Entiendo. De ahora en adelante, usaré solo el inquilino predeterminado (0:0) para todas las operaciones con su instancia de VictoriaMetrics.*\n* > **Tú** : ¿Qué métricas tenemos?\n* > **Asistente** : *Hay muchas métricas disponibles en su instancia de VictoriaMetrics. La lista incluye métricas de varios sistemas y componentes, como:*\n >\n > * *Métricas del sistema ( `node_` , `machine_` , etc.),*\n > * *Métricas de Kubernetes ( `kube_` , `container_` , `apiserver_` , etc.)*\n > * *Métricas específicas de VictoriaMetrics ( `vm_` , `vmagent_` , `vmalert_` , etc.)*\n > * *Métricas de tiempo de ejecución de Go ( `go_` ),*\n > * *Métricas relacionadas con Prometheus ( `scrape_` , `promhttp_` ),*\n > * *Varias métricas de aplicación ( `argocd_` , `alertmanager_` ).*\n >\n > *Esta es una lista completa de más de 2000 métricas disponibles para consulta. ¿Hay alguna métrica específica que le interese explorar más a fondo?*\n* > **Tú** : Bien, a partir de ahora trabajaremos con las métricas de Kubernetes. ¿Podrías decirme cuáles de estas métricas nunca se consultaron?\n* > **Asistente** : *Según los resultados, puedo ver que hay dos métricas de Kubernetes que nunca se han consultado (el recuento de consultas es 0):*\n >\n > * *`kube_persistentvolumeclaim_deletion_timestamp` : esta métrica proporciona información sobre cuándo se marcaron los PersistentVolumeClaims para su eliminación.*\n > * *`kube_pod_init_container_status_last_terminated_reason` : esta métrica muestra el motivo por el cual los contenedores de inicio en los pods se finalizaron por última vez.*\n >\n > *Estas métricas pueden ser valiosas para fines de monitoreo y resolución de problemas:*\n >\n > * *La métrica de marca de tiempo de eliminación de PVC podría ayudar a rastrear patrones de eliminación o identificar PVC atascados.*\n > * *La métrica del motivo de finalización del contenedor de inicialización podría ayudar a identificar patrones de falla comunes en sus contenedores de inicialización.*\n >\n > *¿Quieres que te explique más sobre qué rastrean estas métricas o te sugiera algunas consultas útiles para comenzar a monitorearlas?*\n* > **Tú** : Sí, por favor. Y, además, ¿podrías crear reglas de reetiquetado para dejar de ingerir estas métricas?\n* > **Asistente** : *Crearé una configuración de reetiquetado para excluir de la recopilación estas métricas que nunca se consultan. Esto ayudará a reducir la sobrecarga de almacenamiento y procesamiento. Aquí está la configuración de reetiquetado que puede agregar a su configuración de scraping de VictoriaMetrics:*\n >\n > ```yaml\n > relabel_configs:\n > - action: drop\n > source_labels: [__name__]\n > regex: kube_persistentvolumeclaim_deletion_timestamp|kube_pod_init_container_status_last_terminated_reason\n > ```\n >\n > *Esta configuración hará lo siguiente:*\n >\n > 1. *Mira el nombre de cada métrica (la etiqueta `__name__` )*\n > 2. *Comprueba si coincide con alguna de las métricas especificadas*\n > 3. *Eliminar las métricas que coinciden, evitando que se recopilen y almacenen*\n >\n > *Puede agregar esto a su archivo de configuración de VictoriaMetrics o a las configuraciones de trabajo de scraping relevantes en su configuración de Prometheus si usa Prometheus como scraper para VictoriaMetrics.* *Si desea probar esta configuración de reetiquetado antes de implementarla, puede usar la herramienta de depuración de reetiquetado de métricas para ver cómo afectaría a sus métricas.*\n* > **Tú** : Sí, depura esta configuración de reetiquetado en series de tiempo real para ver cómo afectaría mis métricas.\n* > **Asistente** : *¡Perfecto! La configuración de reetiquetado funciona correctamente.*\n >\n > * *Los primeros dos casos de prueba muestran que las métricas `kube_persistentvolumeclaim_deletion_timestamp` y `kube_pod_init_container_status_last_terminated_reason` se eliminan correctamente (las etiquetas resultantes están vacías `{}` ).*\n > * *El tercer caso de prueba confirma que otras métricas como `kube_pod_status_phase` se conservan y no se ven afectadas por nuestra regla de eliminación.*\n >\n > *Puede agregar con confianza esta configuración de reetiquetado a su configuración de VictoriaMetrics o Prometheus para dejar de recopilar estas métricas no utilizadas. Esto ahorrará espacio de almacenamiento y reducirá la sobrecarga de procesamiento sin afectar sus capacidades de monitorización actuales.*\n* > **Tú** : Pero estoy usando el recurso vmagent del operador VictoriaMetrics, ¿podrías darme parte de las especificaciones para mi agente con esta configuración de reetiquetado?\n* > **Asistente** : *Tienes razón. Dado que estás usando el operador VictoriaMetrics y el recurso personalizado VMAgent, la configuración debe tener el formato adecuado para esa CR. Así es como puedes agregar la configuración de reetiquetado a tu especificación de VMAgent:*\n >\n > ```yaml\n > apiVersion: operator.victoriametrics.com/v1beta1\n > kind: VMAgent\n > metadata:\n > name: your-vmagent-name\n > spec:\n > # ... your existing configuration ...\n > \n > # Add this section to configure global metric relabeling\n > inlineRelabelConfig:\n > - action: drop\n > sourceLabels: [\"__name__\"]\n > regex: kube_persistentvolumeclaim_deletion_timestamp|kube_pod_init_container_status_last_terminated_reason\n > \n > # ... rest of your configuration ...\n > ```\n >\n > *Puntos clave:*\n >\n > * *En la especificación de VMAgent, utilice `inlineRelabelConfig` para las reglas de reetiquetado en línea*\n > * *El campo se llama `sourceLabels` (no `source_labels` como en la configuración de Prometheus)*\n > * *Este es un reetiquetado global que se aplicará a todas las métricas antes de que se almacenen.*\n* > **Tú** : Decidí no eliminar esas métricas, sino simplemente reducir su frecuencia de muestreo mediante la función de submuestreo. ¿Qué parámetros de submuestreo tengo ahora?\n* > **Asistente** : *Puedo ver que actualmente tienes la siguiente configuración de reducción de muestreo:*\n >\n > ```\n > -downsampling.period=\"7d:1m,30d:5m,180d:30m\"\n > ```\n >\n > *Esto significa:*\n >\n > * *Los datos más recientes de 7 días se mantienen en su resolución original.*\n > * *Los datos de entre 7 y 30 días de antigüedad se reducen a una resolución de 1 minuto.*\n > * *Los datos de entre 30 y 180 días de antigüedad se reducen a una resolución de 5 minutos.*\n > * *Los datos con más de 180 días de antigüedad se reducen a una resolución de 30 minutos.*\n >\n > *Ahora, si desea agregar reglas de submuestreo específicas para las métricas que rara vez se consultan, puede usar la herramienta de depuración de filtros de submuestreo para probar el funcionamiento de las nuevas reglas. Probemos una configuración donde aplicamos un submuestreo más agresivo a esas métricas específicas:*\n >\n > *...obteniendo series y usando el depurador de filtro de reducción de muestreo...*\n >\n > *¡Genial! Nuestra prueba ahora funciona correctamente. Basándonos en la configuración actual y en nuestras pruebas, sugerimos una configuración de submuestreo para su CR de VMAgent que aplicaría un submuestreo más agresivo a las métricas que rara vez se consultan:*\n >\n > ```yaml\n > apiVersion: operator.victoriametrics.com/v1beta1\n > kind: VMCluster\n > metadata:\n > name: monitoring-vm-cluster\n > spec:\n > # ... your existing configuration ...\n > \n > # Update the vmstorage section with the new downsampling parameters\n > vmstorage:\n > extraArgs:\n > downsampling.period: '{__name__=\\\"kube_persistentvolumeclaim_deletion_timestamp\\\"}:1d:10m,{__name__=\\\"kube_pod_init_container_status_last_terminated_reason\\\"}:1d:10m,7d:1m,30d:5m,180d:30m'\n > \n > # ... rest of your configuration ...\n > ```\n >\n > *Esta configuración:*\n >\n > *- Aplica un muestreo descendente con una resolución de 10 minutos para las métricas que rara vez se consultan después de solo 1 día.* *- Mantiene su política de muestreo descendente existente para todas las demás métricas.*\n >\n > *El orden de las reglas de reducción de muestreo es importante: las reglas más específicas deben enumerarse primero.*\n\nDurante este diálogo, el asistente estaba utilizando las siguientes herramientas:\n\n* `tenants` para obtener la lista de inquilinos disponibles\n* `documentation` para obtener información sobre la funcionalidad y los formatos de datos\n* `metrics` para obtener la lista de métricas disponibles\n* `metrics_staistics` para obtener información sobre el uso de métricas\n* `series` para obtener la serie temporal para depuración.\n* `metric_relabel_debug` para depurar reglas de reetiquetado\n* `flags` para obtener la información sobre los parámetros de instancia\n* `downsampling_filters_debug` para depurar la configuración de reducción de muestreo\n\nPero puedes utilizar cualquier otra herramienta y combinarlas a tu manera.\n\n## Descargo de responsabilidad\n\nLos servicios y agentes de IA, junto con los servidores MCP como este, no pueden garantizar la precisión, integridad ni fiabilidad de los resultados. Debe verificar los resultados obtenidos con IA. La calidad del servidor MCP y sus respuestas dependen en gran medida de las capacidades de su cliente y de la calidad del modelo que utilice.\n\n## Contribuyendo\n\n¡Agradecemos sus contribuciones al proyecto MCP VictoriaMetrics! No dude en enviar problemas, solicitudes de funcionalidad o solicitudes de incorporación de cambios.","Spanish","es-ES",{"_27":48,"_31":51,"_35":52},{"_29":49,"_31":49,"_32":50},"mcp-ビクトリアメトリックス","# VictoriaMetrics MCP サーバー\n\n[VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/)用の[モデルコンテキストプロトコル (MCP)](https://modelcontextprotocol.io/)サーバーの実装。\n\nこれにより、VictoriaMetricsインスタンスへのアクセスと、 [VictoriaMetrics API](https://docs.victoriametrics.com/victoriametrics/url-examples/)および[ドキュメント](https://docs.victoriametrics.com/)とのシームレスな統合が可能になります。VictoriaMetricsインスタンスに関連する監視、可観測性、デバッグタスクのための包括的なインターフェースを提供し、エンジニアやツールのための高度な自動化とインタラクション機能を実現します。\n\n## 特徴\n\nこの MCP サーバーを使用すると、VictoriaMetrics のほぼすべての読み取り専用 API、つまり[VMUI](https://docs.victoriametrics.com/#vmui)で使用可能なすべての機能を使用できます。\n\n* メトリクスのクエリとデータの探索（クライアントがサポートしている場合はグラフの描画も可能）\n* 利用可能なメトリック、ラベル、ラベル値、およびシリーズ全体の一覧表示とエクスポート\n* アラートと記録ルールとアラートの分析\n* VictoriaMetricsインスタンスのパラメータを表示しています\n* データのカーディナリティとメトリクスの使用統計を調べる\n* クエリの分析\n* 再ラベル付けルール、ダウンサンプリング、保持ポリシー設定のデバッグ\n\nさらに、MCP サーバーには最新のドキュメントが埋め込まれており、オンラインアクセスなしで検索できます。\n\n使用可能なツールとプロンプトの詳細については、 [「使用方法」](#usage)セクションを参照してください。\n\nツールの機能やドキュメント検索機能をプロンプトに組み合わせることで、VictoriaMetricsインスタンスの優れた活用シナリオを構築できます。[ダイアログサンプルの](#dialog-example)セクションで動作例をご確認ください。MCPサーバーとそのレスポンスの品質は、クライアントの能力と使用するモデルの品質に大きく依存することにご注意ください。\n\nMCP サーバーを他の可観測性またはドキュメント検索関連の MCP サーバーと組み合わせて、さらに強力な結果を得ることもできます。\n\n## 要件\n\n* [VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/)インスタンス ([単一ノード](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/)または[クラスター](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/))\n* Go 1.24以上（ソースからビルドする場合）\n\n## インストール\n\n### 行く\n\n```bash\ngo install github.com/VictoriaMetrics-Community/mcp-victoriametrics/cmd/mcp-victoriametrics@latest\n```\n\n### ソースコード\n\n```bash\ngit clone https://github.com/VictoriaMetrics-Community/mcp-victoriametrics.git\ncd mcp-victoriametrics\ngo build -o bin/mcp-victoriametrics ./cmd/mcp-victoriametrics/main.go\n\n# after that add bin/mcp-victoriametrics file to your PATH\n```\n\n### バイナリ\n\n[リリース](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases)ページから最新リリースをダウンロードし、PATH に配置するだけです。\n\n### ドッカー\n\n近日公開...\n\n## 構成\n\nVictoriaMetrics の MCP サーバーは、環境変数を介して構成されます。\n\n| 変数 | 説明 | 必須 | デフォルト | 許容値 |\n| -------------------------- | --------------------------- | --- | ---------------- | -------------------- |\n| `VM_INSTANCE_ENTRYPOINT` | VictoriaMetricsインスタンスへのURL | はい | - | - |\n| `VM_INSTANCE_TYPE` | VictoriaMetricsインスタンスの種類 | はい | - | `single` 、 `cluster` |\n| `VM_INSTANCE_BEARER_TOKEN` | VictoriaMetrics API の認証トークン | いいえ | - | - |\n| `MCP_SERVER_MODE` | サーバー動作モード | いいえ | `stdio` | `stdio` 、 `sse` |\n| `MCP_SSE_ADDR` | SSEサーバがリッスンするアドレス | いいえ | `localhost:8080` | - |\n\n### 設定例\n\n```bash\n# For a single-node instance\nexport VM_INSTANCE_ENTRYPOINT=\"http://localhost:8428\"\nexport VM_INSTANCE_TYPE=\"single\"\nexport VM_INSTANCE_BEARER_TOKEN=\"your-token\"\n\n# For a cluster\nexport VM_INSTANCE_ENTRYPOINT=\"https://play.victoriametrics.com\"\nexport VM_INSTANCE_TYPE=\"cluster\"\n\n# Server mode\nexport MCP_SERVER_MODE=\"sse\"\nexport MCP_SSE_ADDR=\"0.0.0.0:8080\"\n```\n\n## クライアントでの設定\n\n### カーソル\n\n`Settings` -> `Cursor Settings` -> `MCP` -> `Add new global MCP server`に移動し、次の構成をカーソルの`~/.cursor/mcp.json`ファイルに貼り付けます。\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n詳細については、[カーソル MCP のドキュメントを](https://docs.cursor.com/context/model-context-protocol)参照してください。\n\n### クロードデスクトップ\n\nこれを Claude Desktop の`claude_desktop_config.json`ファイルに追加します ( `Settings` -> `Developer` -> `Edit config`を開くと見つかります)。\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n詳細については、 [Claude Desktop MCP のドキュメント](https://modelcontextprotocol.io/quickstart/user)を参照してください。\n\n### クロード・コード\n\n次のコマンドを実行します。\n\n```sh\nclaude mcp add victoriametrics -- /path/to/mcp-victoriametrics \\\n -e VM_INSTANCE_ENTRYPOINT= \\\n -e VM_INSTANCE_TYPE=\n -e VM_INSTANCE_BEARER_TOKEN=\n```\n\n詳細については、 [Claude Code MCP のドキュメント](https://docs.anthropic.com/en/docs/agents-and-tools/claude-code/tutorials#set-up-model-context-protocol-mcp)を参照してください。\n\n### ビジュアルスタジオコード\n\nVS Code MCP 構成ファイルに以下を追加します。\n\n```json\n{\n \"servers\": {\n \"victoriametrics\": {\n \"type\": \"stdio\",\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n詳細については、 [VS Code MCP ドキュメント](https://code.visualstudio.com/docs/copilot/chat/mcp-servers)を参照してください。\n\n### ゼッド\n\nZed 構成ファイルに次の内容を追加します。\n\n```json\n \"context_servers\": {\n \"victoriametrics\": {\n \"command\": {\n \"path\": \"/path/to/mcp-victoriametrics\",\n \"args\": [],\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n },\n \"settings\": {}\n }\n }\n```\n\n詳細については、 [Zed MCP のドキュメント](https://zed.dev/docs/ai/mcp)を参照してください。\n\n### JetBrains IDE\n\n* `Settings` -> `Tools` -> `AI Assistant` -> `Model Context Protocol (MCP)`を開きます。\n* `Add (+)`クリック\n* `As JSON`選択\n* 入力フィールドに次の内容を入力します。\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n### ウィンドサーフィン\n\nWindsurf MCP 構成ファイルに以下を追加します。\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n詳細については、 [Windsurf MCP のドキュメント](https://docs.windsurf.com/windsurf/mcp)を参照してください。\n\n### アマゾンの岩盤\n\n近日公開....\n\n### バイナリの代わりにDockerを使用する\n\n近日公開...\n\n## 使用法\n\nMCP サーバー[をインストールし](#installation)て[構成し](#setup-in-clients)たら、お気に入りの MCP クライアントで使用を開始できます。\n\n次のフレーズから AI アシスタントとの対話を開始できます。\n\n```\nUse MCP VictoriaMetrics in the following answers\n```\n\nしかし、必須ではありません。質問を始めるだけで、アシスタントが自動的にツールとドキュメントを活用して最適な回答を提供します。[ダイアログの例の](#dialog-example)セクションをご覧いただくと、何ができるかをより深く理解できます。\n\n### ツールセット\n\nMCP VictoriaMetrics は、VictoriaMetrics インスタンスと対話するためのさまざまなツールを提供します。\n\n利用可能なツールのリストは次のとおりです。\n\n| 道具 | 説明 |\n| ---------------------------- | -------------------------------------- |\n| `query` | 即時PromQL/MetricsQLクエリを実行する |\n| `query_range` | 一定期間にわたる範囲のPromQL/MetricsQLクエリを実行する |\n| `metrics` | 利用可能な指標の一覧 |\n| `labels` | 利用可能なラベル名を一覧表示する |\n| `label_values` | 特定のラベルの値を一覧表示する |\n| `series` | 利用可能な時系列を一覧表示する |\n| `export` | 生の時系列データをJSONまたはCSVにエクスポートする |\n| `rules` | アラートと記録ルールを表示する |\n| `alerts` | 現在のアラート（発動中および保留中）を表示する |\n| `flags` | VictoriaMetricsインスタンスのデフォルト以外のフラグを表示する |\n| `metric_statistics` | メトリクスの使用状況（クエリ内）の統計情報を取得する |\n| `active_queries` | 現在実行中のクエリを表示する |\n| `top_queries` | 最も頻繁なクエリや最も遅いクエリを表示する |\n| `tsdb_status` | TSDB カーディナリティ統計を表示する |\n| `tenants` | マルチテナントクラスタセットアップで利用可能なテナントを一覧表示する |\n| `documentation` | 埋め込まれたVictoriaMetricsドキュメントを検索 |\n| `metric_relabel_debug` | Prometheus互換の再ラベル付けルールをデバッグする |\n| `downsampling_filters_debug` | ダウンサンプリング構成のデバッグ |\n| `retention_filters_debug` | 保持フィルタの設定をデバッグする |\n\n### プロンプト\n\nサーバーには、一般的なタスク用の事前定義されたプロンプトが含まれています。\n\nこれらは現時点では単なる例であり、プロンプトライブラリは将来追加される予定です。\n\n| プロンプト | 説明 |\n| ------------------------------------------- | -------------------------------------------------- |\n| `unused_metrics` | クエリが実行されないメトリクスを見つけて、それらの取り込みを停止するための再ラベル設定を作成します。 |\n| `documentation` | VictoriaMetricsのドキュメントで特定のトピックを検索 |\n| `rarely_used_metrics_with_high_cardinality` | ほとんどクエリされない高カーディナリティの指標を特定する |\n\n## ダイアログの例\n\nこれは、MCP サーバーを使用したアシスタントとのダイアログの単なる例であり、それがどのように機能するかを理解する上で役立ちます (ただし、これは Victoria Metrics MCP サーバーを Claude Desktop と共に使用し、パブリックプレイ[グラウンドインスタンス](https://play.victoriametrics.com)に構成された実際のダイアログです)。\n\n* > **あなた**: 私の VM ではどのようなテナントが利用できますか?\n* > **アシスタント** *VictoriaMetrics インスタンスでは、次のテナントが利用可能です`0:0` (デフォルトテナント)、 `1:0` `10:0` `100:0`*\n* > **あなた**: 今後はデフォルトのテナントのみを使用してください\n* > **アシスタント**：*了解しました。今後は、VictoriaMetricsインスタンスのすべての操作にデフォルトのテナント（0:0）のみを使用します。*\n* > **あなた**: どのような指標がありますか?\n* > **アシスタント**： *VictoriaMetricsインスタンスでは、多くのメトリクスが利用可能です。リストには、以下のような様々なシステムやコンポーネントからのメトリクスが含まれています。*\n >\n > * *システムメトリック（ `node_` 、 `machine_`など）*\n > * *Kubernetes メトリクス ( `kube_` 、 `container_` 、 `apiserver_`など)*\n > * *VictoriaMetrics固有のメトリック（ `vm_` 、 `vmagent_` 、 `vmalert_`など）*\n > * *Goランタイムメトリクス（ `go_` ）*\n > * *Prometheus関連のメトリクス（ `scrape_` 、 `promhttp_` ）、*\n > * *さまざまなアプリケーションメトリック ( `argocd_` 、 `alertmanager_` )。*\n >\n > *これは、クエリ可能な2,000以上の指標を網羅したリストです。さらに詳しく調べたい指標の種類はありますか？*\n* > **あなた**: では、これからKubernetesのメトリクスを使ってみましょう。これらのメトリクスのうち、クエリが実行されていないものはどれですか？\n* > **アシスタント**:*結果に基づいて、クエリされたことがない Kubernetes メトリックが 2 つあることがわかります (クエリ数は 0)。*\n >\n > * *`kube_persistentvolumeclaim_deletion_timestamp` - このメトリックは、PersistentVolumeClaims が削除対象としてマークされた時期に関する情報を提供します。*\n > * *`kube_pod_init_container_status_last_terminated_reason` - このメトリックは、ポッド内の init コンテナが最後に終了した理由を示します。*\n >\n > *これらのメトリックは、監視やトラブルシューティングに役立つ可能性があります。*\n >\n > * *PVC 削除タイムスタンプメトリックは、削除パターンを追跡したり、スタックされた PVC を識別したりするのに役立ちます。*\n > * *初期化コンテナの終了理由メトリックは、初期化コンテナの一般的な障害パターンを識別するのに役立ちます。*\n >\n > *これらのメトリックが追跡する内容についてさらに詳しく説明したり、監視を開始するための便利なクエリを提案したりしますか?*\n* > **あなた**: はい、お願いします。また、これらのメトリクスの取り込みを停止するための再ラベル付けルールを作成していただけますか？\n* > **アシスタント**：*これらのクエリされていないメトリクスを収集から除外するための再ラベル設定を作成します。これにより、ストレージと処理のオーバーヘッドを削減できます。VictoriaMetricsのスクレイピング設定に追加できる再ラベル設定は次のとおりです。*\n >\n > ```yaml\n > relabel_configs:\n > - action: drop\n > source_labels: [__name__]\n > regex: kube_persistentvolumeclaim_deletion_timestamp|kube_pod_init_container_status_last_terminated_reason\n > ```\n >\n > *この構成では次のようになります。*\n >\n > 1. *各メトリックの名前（ `__name__`ラベル）を確認します。*\n > 2. *指定されたメトリックのいずれかに一致するかどうかを確認します*\n > 3. *一致するメトリックを削除し、収集および保存されないようにします。*\n >\n > *VictoriaMetricsの設定ファイルに追加するか、PrometheusをVictoriaMetricsのスクレイパーとして使用している場合は、Prometheus設定内の関連するスクレイピングジョブ設定に追加してください。* *この再ラベル設定を実装前にテストしたい場合は、メトリック再ラベルデバッグツールを使用して、メトリックにどのような影響があるかを確認できます。*\n* > **あなた**: はい、このラベル変更構成をリアルタイムシリーズでデバッグして、それが私のメトリクスにどのような影響を与えるかを確認してください。\n* > **アシスタント**:*完璧です! ラベル変更の設定は期待通りに動作しました:*\n >\n > * *最初の 2 つのテストケースでは、 `kube_persistentvolumeclaim_deletion_timestamp`と`kube_pod_init_container_status_last_terminated_reason`両方のメトリックが正しくドロップされていることがわかります (結果のラベルは空の`{}`になります)。*\n > * *3 番目のテストケースでは、 `kube_pod_status_phase`などの他のメトリックが保持され、ドロップルールの影響を受けないことを確認します。*\n >\n > *VictoriaMetricsまたはPrometheusの設定にこの再ラベル設定を追加することで、これらの未使用のメトリクスの収集を停止できます。これにより、現在の監視機能に影響を与えることなく、ストレージ容量を節約し、処理オーバーヘッドを削減できます。*\n* > **あなた**: しかし、私は VictoriaMetrics オペレーターの vmagent リソースを使用しているので、この再ラベル構成を使用したエージェントの仕様の一部を教えていただけますか?\n* > **アシスタント**:*おっしゃる通りです。VictoriaMetrics Operator と VMAgent カスタムリソースをご利用なので、設定は CR に適した形式である必要があります。VMAgent 仕様に relabel 設定を追加する手順は以下のとおりです。*\n >\n > ```yaml\n > apiVersion: operator.victoriametrics.com/v1beta1\n > kind: VMAgent\n > metadata:\n > name: your-vmagent-name\n > spec:\n > # ... your existing configuration ...\n > \n > # Add this section to configure global metric relabeling\n > inlineRelabelConfig:\n > - action: drop\n > sourceLabels: [\"__name__\"]\n > regex: kube_persistentvolumeclaim_deletion_timestamp|kube_pod_init_container_status_last_terminated_reason\n > \n > # ... rest of your configuration ...\n > ```\n >\n > *要点:*\n >\n > * *VMAgent仕様では、インライン再ラベル付けルールに`inlineRelabelConfig`を使用します。*\n > * *このフィールドは`sourceLabels`と呼ばれます（Prometheus設定の`source_labels`ではありません）。*\n > * *これは、保存される前にすべての指標に適用されるグローバルな再ラベル付けです。*\n* > **あなた**: これらのメトリックは削除せず、ダウンサンプリングを使ってサンプルレートを下げることにしました。ダウンサンプリングのパラメータは今どうなっていますか？\n* > **アシスタント**:*現在、ダウンサンプリング構成は次のようになっています。*\n >\n > ```\n > -downsampling.period=\"7d:1m,30d:5m,180d:30m\"\n > ```\n >\n > *これはつまり：*\n >\n > * *7日以上経過したデータは元の解像度で保存されます*\n > * *7～30日前のデータは1分単位の解像度にダウンサンプリングされます*\n > * *30～180日前のデータは5分単位の解像度にダウンサンプリングされます*\n > * *180日以上前のデータは30分単位にダウンサンプリングされます*\n >\n > *クエリ頻度の低い指標に特定のダウンサンプリングルールを追加したい場合は、ダウンサンプリングフィルターのデバッグツールを使用して、新しいルールの動作をテストできます。これらの特定の指標に、より積極的なダウンサンプリングを適用する設定をテストしてみましょう。*\n >\n > *...シリーズを取得し、ダウンサンプリングフィルターデバッガーを使用しています...*\n >\n > *素晴らしい！これでテストは正常に動作しています。現在の設定とテスト結果に基づき、VMAgent CR の推奨ダウンサンプリング設定を以下に示します。この設定では、クエリ頻度の低いメトリクスに対して、より積極的なダウンサンプリングが適用されます。*\n >\n > ```yaml\n > apiVersion: operator.victoriametrics.com/v1beta1\n > kind: VMCluster\n > metadata:\n > name: monitoring-vm-cluster\n > spec:\n > # ... your existing configuration ...\n > \n > # Update the vmstorage section with the new downsampling parameters\n > vmstorage:\n > extraArgs:\n > downsampling.period: '{__name__=\\\"kube_persistentvolumeclaim_deletion_timestamp\\\"}:1d:10m,{__name__=\\\"kube_pod_init_container_status_last_terminated_reason\\\"}:1d:10m,7d:1m,30d:5m,180d:30m'\n > \n > # ... rest of your configuration ...\n > ```\n >\n > *この構成:*\n >\n > *- わずか 1 日後に、ほとんどクエリされないメトリクスに対して 10 分解像度のダウンサンプリングを適用します* *- 他のすべてのメトリクスに対しては、既存のダウンサンプリングポリシーを維持します*\n >\n > *ダウンサンプリングルールの順序は重要です。より具体的なルールを最初にリストする必要があります。*\n\nこのダイアログ中、アシスタントは次のツールを使用していました。\n\n* 利用可能なテナントのリストを取得するには、 `tenants`使用します。\n* 機能とデータ形式に関する情報を取得するための`documentation`\n* 利用可能なメトリックのリストを取得するための`metrics`\n* メトリクスの使用状況に関する情報を取得するための`metrics_staistics`\n* デバッグ用の時系列を取得するための`series`\n* 再ラベル付けルールをデバッグするための`metric_relabel_debug`\n* インスタンスパラメータに関する情報を取得するための`flags`\n* ダウンサンプリング設定をデバッグするための`downsampling_filters_debug`\n\nただし、他のツールを使用して、独自の方法で組み合わせることもできます。\n\n## 免責事項\n\nAIサービスやエージェント、そしてこのようなMCPサーバーは、結果の正確性、完全性、信頼性を保証するものではありません。AIで得られた結果は、ご自身で再確認する必要があります。MCPサーバーとその応答の品質は、クライアントの能力と使用するモデルの品質に大きく依存します。\n\n## 貢献\n\nMCP VictoriaMetrics プロジェクトへの貢献を歓迎します。問題、機能リクエスト、プルリクエストなど、お気軽にご提出ください。","Japanese","ja-JP",{"_27":54,"_31":56,"_35":57},{"_29":30,"_31":30,"_32":55},"# VictoriaMetrics MCP 서버\n\n[VictoriaMetrics를](https://docs.victoriametrics.com/victoriametrics/) 위한 [MCP(Model Context Protocol)](https://modelcontextprotocol.io/) 서버 구현.\n\n이를 통해 VictoriaMetrics 인스턴스에 대한 접근 권한과 [VictoriaMetrics API](https://docs.victoriametrics.com/victoriametrics/url-examples/) 및 [문서](https://docs.victoriametrics.com/) 와의 원활한 통합이 제공됩니다. VictoriaMetrics 인스턴스 관련 모니터링, 관측 및 디버깅 작업을 위한 포괄적인 인터페이스를 제공하고, 엔지니어와 도구를 위한 고급 자동화 및 상호 작용 기능을 활성화할 수 있습니다.\n\n## 특징\n\n이 MCP 서버를 사용하면 VictoriaMetrics의 거의 모든 읽기 전용 API, 즉 [VMUI](https://docs.victoriametrics.com/#vmui) 에서 사용 가능한 모든 기능을 사용할 수 있습니다.\n\n* 메트릭 쿼리 및 데이터 탐색(클라이언트가 지원하는 경우 그래프 그리기도 가능)\n* 사용 가능한 메트릭, 레이블, 레이블 값 및 전체 시리즈 나열 및 내보내기\n* 알림 및 녹음 규칙과 알림 분석\n* VictoriaMetrics 인스턴스의 매개변수 표시\n* 데이터 및 메트릭 사용 통계의 기수 탐색\n* 귀하의 쿼리 분석\n* 재레이블링 규칙, 다운샘플링 및 보존 정책 구성 디버깅\n\n또한 MCP 서버에는 최신 문서가 내장되어 있어 온라인 접속 없이도 검색할 수 있습니다.\n\n사용 가능한 도구와 프롬프트에 대한 자세한 내용은 [사용](#usage) 섹션에서 확인할 수 있습니다.\n\n도구 기능, 문서 검색 기능을 프롬프트에 결합하여 VictoriaMetrics 인스턴스에 대한 훌륭한 사용 시나리오를 만들어낼 수 있습니다. 작동 방식은 [Dialog 예제](#dialog-example) 섹션을 참조하세요. MCP 서버와 응답의 품질은 클라이언트의 성능과 사용 중인 모델의 품질에 따라 크게 달라진다는 점에 유의하세요.\n\nMCP 서버를 다른 관찰성 또는 문서 검색 관련 MCP 서버와 결합하여 더욱 강력한 결과를 얻을 수도 있습니다.\n\n## 요구 사항\n\n* [VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/) 인스턴스( [단일 노드](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/) 또는 [클러스터](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/) )\n* Go 1.24 이상(소스에서 빌드하려는 경우)\n\n## 설치\n\n### 가다\n\n지엑스피1\n\n### 소스 코드\n\n```bash\ngit clone https://github.com/VictoriaMetrics-Community/mcp-victoriametrics.git\ncd mcp-victoriametrics\ngo build -o bin/mcp-victoriametrics ./cmd/mcp-victoriametrics/main.go\n\n# after that add bin/mcp-victoriametrics file to your PATH\n```\n\n### 바이너리\n\n[릴리스](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases) 페이지에서 최신 릴리스를 다운로드하고 PATH에 추가하세요.\n\n### 도커\n\n곧 출시됩니다...\n\n## 구성\n\nVictoriaMetrics용 MCP 서버는 환경 변수를 통해 구성됩니다.\n\n| 변하기 쉬운 | 설명 | 필수의 | 기본 | 허용된 값 |\n| -------------------------- | ----------------------------- | --- | ---------------- | -------------------- |\n| `VM_INSTANCE_ENTRYPOINT` | VictoriaMetrics 인스턴스의 URL | 예 | - | - |\n| `VM_INSTANCE_TYPE` | VictoriaMetrics 인스턴스 유형 | 예 | - | `single` , `cluster` |\n| `VM_INSTANCE_BEARER_TOKEN` | VictoriaMetrics API에 대한 인증 토큰 | 아니요 | - | - |\n| `MCP_SERVER_MODE` | 서버 작동 모드 | 아니요 | `stdio` | `stdio` , `sse` |\n| `MCP_SSE_ADDR` | SSE 서버가 수신할 주소 | 아니요 | `localhost:8080` | - |\n\n### 구성 예\n\n```bash\n# For a single-node instance\nexport VM_INSTANCE_ENTRYPOINT=\"http://localhost:8428\"\nexport VM_INSTANCE_TYPE=\"single\"\nexport VM_INSTANCE_BEARER_TOKEN=\"your-token\"\n\n# For a cluster\nexport VM_INSTANCE_ENTRYPOINT=\"https://play.victoriametrics.com\"\nexport VM_INSTANCE_TYPE=\"cluster\"\n\n# Server mode\nexport MCP_SERVER_MODE=\"sse\"\nexport MCP_SSE_ADDR=\"0.0.0.0:8080\"\n```\n\n## 클라이언트 설정\n\n### 커서\n\n`Settings` -> `Cursor Settings` -> `MCP` -> `Add new global MCP server` 로 이동하여 다음 구성을 커서 `~/.cursor/mcp.json` 파일에 붙여넣습니다.\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n자세한 내용은 [커서 MCP 문서를](https://docs.cursor.com/context/model-context-protocol) 참조하세요.\n\n### 클로드 데스크탑\n\nClaude Desktop `claude_desktop_config.json` 파일에 다음을 추가합니다( `Settings` -> `Developer` -> `Edit config` 열면 찾을 수 있습니다).\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n자세한 내용은 [Claude Desktop MCP 문서를](https://modelcontextprotocol.io/quickstart/user) 참조하세요.\n\n### 클로드 코드\n\n다음 명령을 실행합니다.\n\n```sh\nclaude mcp add victoriametrics -- /path/to/mcp-victoriametrics \\\n -e VM_INSTANCE_ENTRYPOINT= \\\n -e VM_INSTANCE_TYPE=\n -e VM_INSTANCE_BEARER_TOKEN=\n```\n\n자세한 내용은 [Claude Code MCP 문서를](https://docs.anthropic.com/en/docs/agents-and-tools/claude-code/tutorials#set-up-model-context-protocol-mcp) 참조하세요.\n\n### 비주얼 스튜디오 코드\n\nVS Code MCP 구성 파일에 다음을 추가하세요.\n\n```json\n{\n \"servers\": {\n \"victoriametrics\": {\n \"type\": \"stdio\",\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n자세한 내용은 [VS Code MCP 문서를](https://code.visualstudio.com/docs/copilot/chat/mcp-servers) 참조하세요.\n\n### 제드 자의 명칭\n\nZed 구성 파일에 다음을 추가하세요.\n\n```json\n \"context_servers\": {\n \"victoriametrics\": {\n \"command\": {\n \"path\": \"/path/to/mcp-victoriametrics\",\n \"args\": [],\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n },\n \"settings\": {}\n }\n }\n```\n\n자세한 내용은 [Zed MCP 문서를](https://zed.dev/docs/ai/mcp) 참조하세요.\n\n### JetBrains IDE\n\n* `Settings` -> `Tools` -> `AI Assistant` -> `Model Context Protocol (MCP)` 엽니다.\n* `Add (+)` 클릭하세요\n* `As JSON` 선택\n* 입력 필드에 다음을 입력하세요.\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n### 윈드서핑\n\nWindsurf MCP 구성 파일에 다음을 추가하세요.\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n자세한 내용은 [Windsurf MCP 문서를](https://docs.windsurf.com/windsurf/mcp) 참조하세요.\n\n### 아마존 베드락\n\n곧 출시됩니다....\n\n### 바이너리 대신 Docker 사용\n\n곧 출시됩니다...\n\n## 용법\n\nMCP 서버를 [설치](#installation) 하고 [구성한](#setup-in-clients) 후에는 원하는 MCP 클라이언트와 함께 사용할 수 있습니다.\n\n다음 문구를 통해 AI 도우미와 대화를 시작할 수 있습니다.\n\n```\nUse MCP VictoriaMetrics in the following answers\n```\n\n하지만 필수는 아닙니다. 질문을 하기만 하면 어시스턴트가 자동으로 도구와 문서를 활용하여 최상의 답변을 제공합니다. [Dialog 예시](#dialog-example) 섹션을 참고하시면 어떤 기능을 사용할 수 있는지 더 자세히 이해하실 수 있습니다.\n\n### 도구 세트\n\nMCP VictoriaMetrics는 VictoriaMetrics 인스턴스와 상호 작용하기 위한 다양한 도구를 제공합니다.\n\n사용 가능한 도구 목록은 다음과 같습니다.\n\n| 도구 | 설명 |\n| ---------------------------- | ----------------------------------- |\n| `query` | PromQL/MetricsQL 쿼리를 즉시 실행합니다. |\n| `query_range` | 특정 기간 동안 PromQL/MetricsQL 쿼리 범위 실행 |\n| `metrics` | 사용 가능한 메트릭 나열 |\n| `labels` | 사용 가능한 레이블 이름 나열 |\n| `label_values` | 특정 레이블에 대한 값 나열 |\n| `series` | 사용 가능한 시계열 목록 |\n| `export` | 원시 시계열 데이터를 JSON 또는 CSV로 내보내기 |\n| `rules` | 알림 및 녹음 규칙 보기 |\n| `alerts` | 현재 알림(발사 및 보류) 보기 |\n| `flags` | VictoriaMetrics 인스턴스의 기본이 아닌 플래그 보기 |\n| `metric_statistics` | (쿼리에서) 메트릭 사용 통계 가져오기 |\n| `active_queries` | 현재 실행 중인 쿼리 보기 |\n| `top_queries` | 가장 빈번하거나 가장 느린 쿼리 보기 |\n| `tsdb_status` | TSDB 기수 통계 보기 |\n| `tenants` | 멀티 테넌트 클러스터 설정에서 사용 가능한 테넌트 나열 |\n| `documentation` | 내장된 VictoriaMetrics 문서에서 검색 |\n| `metric_relabel_debug` | Prometheus 호환 재레이블링 규칙 디버그 |\n| `downsampling_filters_debug` | 디버그 다운샘플링 구성 |\n| `retention_filters_debug` | 디버그 보존 필터 구성 |\n\n### 프롬프트\n\n서버에는 일반적인 작업에 대한 미리 정의된 프롬프트가 포함되어 있습니다.\n\n현재는 단지 예시일 뿐이며, 프롬프트 라이브러리는 앞으로 추가될 예정입니다.\n\n| 즉각적인 | 설명 |\n| ------------------------------------------- | -------------------------------------------- |\n| `unused_metrics` | 쿼리되지 않는 메트릭을 찾아 이를 수집하지 못하도록 재레이블 구성을 생성합니다. |\n| `documentation` | 특정 주제에 대한 VictoriaMetrics 문서 검색 |\n| `rarely_used_metrics_with_high_cardinality` | 거의 쿼리되지 않는 높은 카디널리티 메트릭 식별 |\n\n## 대화 예제\n\n이는 MCP 서버를 사용하는 Assistant와의 대화의 예일 뿐이며, 작동 방식을 이해하는 데 도움이 될 수 있습니다(하지만 실제로는 Claude Desktop과 [공개 플레이그라운드 인스턴스](https://play.victoriametrics.com) 로 구성된 Victoria Metrics MCP 서버를 사용하는 대화입니다).\n\n* > **귀하** : 내 VM에서 어떤 테넌트를 사용할 수 있나요?\n* > *VictoriaMetrics 인스턴스에서 다음 테넌트를 사용할 수 있습니다. `0:0` (기본 테넌트), `1:0` , `10:0` , `100:0`* **0**\n* > **귀하** : 지금부터는 기본 테넌트만 사용하세요.\n* > **Assistant** : *알겠습니다. 이제부터 VictoriaMetrics 인스턴스 관련 모든 작업에 기본 테넌트(0:0)만 사용하겠습니다.*\n* > **당신** : 우리는 어떤 측정 항목을 가지고 있나요?\n* > **Assistant** : *VictoriaMetrics 인스턴스에는 다양한 메트릭이 있습니다. 이 목록에는 다음과 같은 다양한 시스템 및 구성 요소의 메트릭이 포함됩니다.*\n >\n > * *시스템 메트릭( `node_` , `machine_` 등)*\n > * *Kubernetes 메트릭( `kube_` , `container_` , `apiserver_` 등)*\n > * *VictoriaMetrics 특정 메트릭( `vm_` , `vmagent_` , `vmalert_` 등)*\n > * *Go 런타임 메트릭( `go_` ),*\n > * *Prometheus 관련 메트릭( `scrape_` , `promhttp_` ),*\n > * *다양한 애플리케이션 메트릭( `argocd_` , `alertmanager_` ).*\n >\n > *2,000개 이상의 지표를 포괄적으로 정리한 목록입니다. 더 자세히 살펴보고 싶은 특정 유형의 지표가 있으신가요?*\n* > **당신** : 좋아요, 이제부터 쿠버네티스 메트릭을 사용해 봅시다. 이 메트릭 중 쿼리되지 않은 메트릭이 무엇인지 알려주세요.\n* > **Assistant** : *결과를 바탕으로, 쿼리된 적이 없는 Kubernetes 메트릭이 두 개 있는 것을 알 수 있습니다(쿼리 수는 0입니다).*\n >\n > * *`kube_persistentvolumeclaim_deletion_timestamp` - 이 메트릭은 PersistentVolumeClaims가 삭제 표시된 시점에 대한 정보를 제공합니다.*\n > * *`kube_pod_init_container_status_last_terminated_reason` - 이 메트릭은 Pod 내 init 컨테이너가 마지막으로 종료된 이유를 보여줍니다.*\n >\n > *이러한 측정항목은 모니터링 및 문제 해결 목적으로 유용할 수 있습니다.*\n >\n > * *PVC 삭제 타임스탬프 메트릭은 삭제 패턴을 추적하거나 중단된 PVC를 식별하는 데 도움이 될 수 있습니다.*\n > * *init 컨테이너 종료 이유 메트릭은 초기화 컨테이너에서 일반적인 실패 패턴을 식별하는 데 도움이 될 수 있습니다.*\n >\n > *이러한 지표가 추적하는 내용에 대해 자세히 설명해 주시겠습니까? 아니면 이를 모니터링하기 위한 유용한 쿼리를 제안해 주시겠습니까?*\n* > **당신** : 네, 부탁드립니다. 그리고, 이러한 지표 수집을 막기 위해 라벨 재지정 규칙을 만들어 주실 수 있나요?\n* > **Assistant** : *쿼리되지 않은 이러한 메트릭을 수집에서 제외하는 재라벨링 구성을 생성하겠습니다. 이렇게 하면 저장 및 처리 오버헤드를 줄이는 데 도움이 됩니다. VictoriaMetrics 스크레이핑 구성에 추가할 수 있는 재라벨링 구성은 다음과 같습니다.*\n >\n > ```yaml\n > relabel_configs:\n > - action: drop\n > source_labels: [__name__]\n > regex: kube_persistentvolumeclaim_deletion_timestamp|kube_pod_init_container_status_last_terminated_reason\n > ```\n >\n > *이 구성은 다음을 수행합니다.*\n >\n > 1. *각 메트릭의 이름( `__name__` 레이블)을 살펴보세요.*\n > 2. *지정된 메트릭 중 하나와 일치하는지 확인하세요.*\n > 3. *일치하는 메트릭을 삭제하여 수집 및 저장을 방지합니다.*\n >\n > *VictoriaMetrics의 스크래퍼로 Prometheus를 사용하는 경우, VictoriaMetrics 구성 파일이나 Prometheus 구성의 관련 스크래핑 작업 구성에 이 설정을 추가할 수 있습니다.* *이 재레이블 구성을 구현하기 전에 테스트하려면 메트릭 재레이블 디버그 도구를 사용하여 메트릭에 어떤 영향을 미치는지 확인할 수 있습니다.*\n* > **귀하** : 예, 실시간 시리즈에서 이 재라벨 구성을 디버깅하여 내 메트릭에 어떤 영향을 미치는지 확인하십시오.\n* > **Assistant** : *완벽해요! 라벨 재지정 구성이 예상대로 작동합니다.*\n >\n > * *처음 두 테스트 사례는 `kube_persistentvolumeclaim_deletion_timestamp` 와 `kube_pod_init_container_status_last_terminated_reason` 메트릭이 모두 올바르게 삭제되었음을 보여줍니다(결과 레이블은 비어 있음 `{}` ).*\n > * *세 번째 테스트 사례에서는 `kube_pod_status_phase` 와 같은 다른 메트릭이 보존되고 삭제 규칙의 영향을 받지 않는다는 것을 확인합니다.*\n >\n > *VictoriaMetrics 또는 Prometheus 구성에 이 재레이블 구성을 추가하면 사용하지 않는 메트릭 수집을 중단할 수 있습니다. 이렇게 하면 현재 모니터링 기능에 영향을 주지 않고 저장 공간을 절약하고 처리 오버헤드를 줄일 수 있습니다.*\n* > **당신** : 하지만 저는 VictoriaMetrics 연산자의 vmagent 리소스를 사용하고 있는데, 이 재레이블 구성이 적용된 에이전트에 대한 사양의 일부를 알려주시겠습니까?\n* > **Assistant** : *맞습니다. VictoriaMetrics Operator와 VMAgent 사용자 지정 리소스를 사용하고 있으므로 구성은 해당 CR에 적합한 형식이어야 합니다. VMAgent 사양에 레이블 재지정 구성을 추가하는 방법은 다음과 같습니다.*\n >\n > ```yaml\n > apiVersion: operator.victoriametrics.com/v1beta1\n > kind: VMAgent\n > metadata:\n > name: your-vmagent-name\n > spec:\n > # ... your existing configuration ...\n > \n > # Add this section to configure global metric relabeling\n > inlineRelabelConfig:\n > - action: drop\n > sourceLabels: [\"__name__\"]\n > regex: kube_persistentvolumeclaim_deletion_timestamp|kube_pod_init_container_status_last_terminated_reason\n > \n > # ... rest of your configuration ...\n > ```\n >\n > *핵심 포인트:*\n >\n > * *VMAgent 사양에서는 인라인 재레이블링 규칙에 대해 `inlineRelabelConfig` 사용합니다.*\n > * *필드는 `sourceLabels` 라고 합니다(Prometheus 구성의 `source_labels` 가 아님)*\n > * *이는 모든 메트릭이 저장되기 전에 적용되는 글로벌 재레이블링입니다.*\n* > **당신** : 해당 지표를 제거하지 않고 다운샘플링을 사용하여 샘플링 레이트만 낮추기로 했습니다. 현재 제가 가진 다운샘플링 매개변수는 무엇인가요?\n* > **보조원** : *현재 다음과 같은 다운샘플링 구성이 있는 것으로 보입니다.*\n >\n > ```\n > -downsampling.period=\"7d:1m,30d:5m,180d:30m\"\n > ```\n >\n > *이는 다음을 의미합니다.*\n >\n > * *7일 이상 지난 데이터는 원래 해상도로 보관됩니다.*\n > * *7~30일 전 데이터는 1분 해상도로 다운샘플링됩니다.*\n > * *30~180일 된 데이터는 5분 해상도로 다운샘플링됩니다.*\n > * *180일 이상 된 데이터는 30분 해상도로 다운샘플링됩니다.*\n >\n > *이제 쿼리 빈도가 낮은 지표에 대해 특정 다운샘플링 규칙을 추가하려면 다운샘플링 필터 디버그 도구를 사용하여 새 규칙이 어떻게 작동하는지 테스트할 수 있습니다. 특정 지표에 더욱 공격적인 다운샘플링을 적용하는 구성을 테스트해 보겠습니다.*\n >\n > *...시리즈를 가져오고 다운샘플링 필터 디버거를 사용합니다...*\n >\n > *좋습니다! 이제 테스트가 정상적으로 작동합니다. 현재 구성과 테스트 결과를 바탕으로, VMAgent CR에 대해 쿼리 빈도가 낮은 메트릭에 더욱 적극적인 다운샘플링을 적용하는 다운샘플링 구성을 제안합니다.*\n >\n > ```yaml\n > apiVersion: operator.victoriametrics.com/v1beta1\n > kind: VMCluster\n > metadata:\n > name: monitoring-vm-cluster\n > spec:\n > # ... your existing configuration ...\n > \n > # Update the vmstorage section with the new downsampling parameters\n > vmstorage:\n > extraArgs:\n > downsampling.period: '{__name__=\\\"kube_persistentvolumeclaim_deletion_timestamp\\\"}:1d:10m,{__name__=\\\"kube_pod_init_container_status_last_terminated_reason\\\"}:1d:10m,7d:1m,30d:5m,180d:30m'\n > \n > # ... rest of your configuration ...\n > ```\n >\n > *이 구성:*\n >\n > *- 거의 쿼리되지 않는 메트릭에 대해 단 1일 만에 10분 해상도 다운샘플링을 적용합니다.* *- 다른 모든 메트릭에 대해 기존 다운샘플링 정책을 유지합니다.*\n >\n > *다운샘플링 규칙의 순서는 중요합니다. 보다 구체적인 규칙은 먼저 나열해야 합니다.*\n\n이 대화 중에 보조자는 다음 도구를 사용했습니다.\n\n* 사용 가능한 테넌트 목록을 얻으려면 `tenants` .\n* 기능 및 데이터 형식에 대한 정보를 얻기 위한 `documentation`\n* 사용 가능한 메트릭 목록을 가져오려면 `metrics`\n* `metrics_staistics` 사용하여 메트릭 사용에 대한 정보를 얻습니다.\n* 디버깅을 위한 시계열을 얻기 위한 `series`\n* `metric_relabel_debug` 사용하여 재레이블링 규칙을 디버깅합니다.\n* 인스턴스 매개변수에 대한 정보를 얻기 위한 `flags`\n* 다운샘플링 구성을 디버깅하기 위한 `downsampling_filters_debug`\n\n하지만 다른 도구를 사용하고 원하는 대로 결합할 수도 있습니다.\n\n## 부인 성명\n\n이러한 MCP 서버와 함께 제공되는 AI 서비스 및 에이전트는 결과의 정확성, 완전성 및 신뢰성을 보장할 수 없습니다. AI를 통해 얻은 결과를 다시 한번 확인해야 합니다. MCP 서버와 그 응답의 품질은 클라이언트의 성능과 사용 중인 모델의 품질에 따라 크게 달라집니다.\n\n## 기여하다\n\nMCP VictoriaMetrics 프로젝트에 대한 기여를 환영합니다! 이슈, 기능 요청 또는 풀 리퀘스트를 자유롭게 제출해 주세요.","Korean","ko-KR",{"_27":59,"_31":61,"_35":62},{"_29":30,"_31":30,"_32":60},"# Сервер MCP VictoriaMetrics\n\nРеализация сервера [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) для [VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/) .\n\nЭто обеспечивает доступ к вашему экземпляру VictoriaMetrics и бесшовную интеграцию с API и [документацией](https://docs.victoriametrics.com/) [VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/url-examples/) . Это может предоставить вам комплексный интерфейс для мониторинга, наблюдения и отладки задач, связанных с вашими экземплярами VictoriaMetrics, включить расширенные возможности автоматизации и взаимодействия для инженеров и инструментов.\n\n## Функции\n\nЭтот сервер MCP позволяет использовать почти все API VictoriaMetrics, доступные только для чтения, т. е. все функции, доступные в [VMUI](https://docs.victoriametrics.com/#vmui) :\n\n* Запрос показателей и исследование данных (даже построение графиков, если ваш клиент это поддерживает)\n* Перечисление и экспорт доступных метрик, меток, значений меток и целых серий\n* Анализ правил оповещения и регистрации, а также оповещений\n* Отображение параметров вашего экземпляра VictoriaMetrics\n* Изучение мощности ваших данных и статистики использования метрик\n* Анализ ваших запросов\n* Отладка правил перемаркировки, понижения выборки и конфигураций политик хранения\n\nКроме того, сервер MCP содержит встроенную актуальную документацию и позволяет осуществлять поиск по ней без доступа в Интернет.\n\nБолее подробную информацию о доступных инструментах и подсказках можно найти в разделе [«Использование»](#usage) .\n\nВы можете объединить функциональность инструментов, поиска документов в ваших подсказках и придумать отличные сценарии использования для вашего экземпляра VictoriaMetrics. Просто проверьте раздел [Пример диалога,](#dialog-example) чтобы увидеть, как это может работать. И, пожалуйста, обратите внимание на тот факт, что качество MCP Server и его ответов во многом зависит от возможностей вашего клиента и качества модели, которую вы используете.\n\nВы также можете объединить сервер MCP с другими серверами MCP, связанными с наблюдением или поиском документов, и получить еще более мощные результаты.\n\n## Требования\n\n* Экземпляр [VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/) ( [одноузловой](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/) или [кластерный](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/) )\n* Используйте версию 1.24 или выше (если вы хотите собрать из исходников)\n\n## Установка\n\n### Идти\n\n```bash\ngo install github.com/VictoriaMetrics-Community/mcp-victoriametrics/cmd/mcp-victoriametrics@latest\n```\n\n### Исходный код\n\n```bash\ngit clone https://github.com/VictoriaMetrics-Community/mcp-victoriametrics.git\ncd mcp-victoriametrics\ngo build -o bin/mcp-victoriametrics ./cmd/mcp-victoriametrics/main.go\n\n# after that add bin/mcp-victoriametrics file to your PATH\n```\n\n### Двоичные файлы\n\nПросто загрузите последнюю версию со страницы [«Релизы»](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases) и добавьте ее в PATH.\n\n### Докер\n\nВскоре...\n\n## Конфигурация\n\nСервер MCP для VictoriaMetrics настраивается с помощью переменных среды:\n\n| Переменная | Описание | Необходимый | По умолчанию | Допустимые значения |\n| -------------------------- | -------------------------------------------- | ----------- | ---------------- | -------------------- |\n| `VM_INSTANCE_ENTRYPOINT` | URL-адрес экземпляра VictoriaMetrics | Да | - | - |\n| `VM_INSTANCE_TYPE` | Тип экземпляра VictoriaMetrics | Да | - | `single` , `cluster` |\n| `VM_INSTANCE_BEARER_TOKEN` | Токен аутентификации для API VictoriaMetrics | Нет | - | - |\n| `MCP_SERVER_MODE` | Режим работы сервера | Нет | `stdio` | `stdio` , `sse` |\n| `MCP_SSE_ADDR` | Адрес для прослушивания сервером SSE | Нет | `localhost:8080` | - |\n\n### Примеры конфигурации\n\n```bash\n# For a single-node instance\nexport VM_INSTANCE_ENTRYPOINT=\"http://localhost:8428\"\nexport VM_INSTANCE_TYPE=\"single\"\nexport VM_INSTANCE_BEARER_TOKEN=\"your-token\"\n\n# For a cluster\nexport VM_INSTANCE_ENTRYPOINT=\"https://play.victoriametrics.com\"\nexport VM_INSTANCE_TYPE=\"cluster\"\n\n# Server mode\nexport MCP_SERVER_MODE=\"sse\"\nexport MCP_SSE_ADDR=\"0.0.0.0:8080\"\n```\n\n## Настройка в клиентах\n\n### Курсор\n\nПерейдите в: `Settings` -> `Cursor Settings` -> `MCP` -> `Add new global MCP server` и вставьте следующую конфигурацию в файл курсора `~/.cursor/mcp.json` :\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\nБолее подробную информацию см. [в документации Cursor MCP](https://docs.cursor.com/context/model-context-protocol) .\n\n### Клод Десктоп\n\nДобавьте это в файл `claude_desktop_config.json` вашего рабочего стола Claude (его можно найти, открыв `Settings` -> `Developer` -> `Edit config` ):\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\nБолее подробную информацию см. [в документации Claude Desktop MCP](https://modelcontextprotocol.io/quickstart/user) .\n\n### Клод Код\n\nВыполните команду:\n\n```sh\nclaude mcp add victoriametrics -- /path/to/mcp-victoriametrics \\\n -e VM_INSTANCE_ENTRYPOINT= \\\n -e VM_INSTANCE_TYPE=\n -e VM_INSTANCE_BEARER_TOKEN=\n```\n\nБолее подробную информацию см. [в документации Claude Code MCP](https://docs.anthropic.com/en/docs/agents-and-tools/claude-code/tutorials#set-up-model-context-protocol-mcp) .\n\n### Код Visual Studio\n\nДобавьте это в файл конфигурации VS Code MCP:\n\n```json\n{\n \"servers\": {\n \"victoriametrics\": {\n \"type\": \"stdio\",\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\nБолее подробную информацию см. [в документации VS Code MCP](https://code.visualstudio.com/docs/copilot/chat/mcp-servers) .\n\n### Зед\n\nДобавьте следующее в файл конфигурации Zed:\n\n```json\n \"context_servers\": {\n \"victoriametrics\": {\n \"command\": {\n \"path\": \"/path/to/mcp-victoriametrics\",\n \"args\": [],\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n },\n \"settings\": {}\n }\n }\n```\n\nБолее подробную информацию см. [в документации Zed MCP](https://zed.dev/docs/ai/mcp) .\n\n### JetBrains IDE\n\n* Откройте `Settings` -> `Tools` -> `AI Assistant` -> `Model Context Protocol (MCP)` .\n* Нажмите `Add (+)`\n* Выбрать `As JSON`\n* Введите в поле ввода следующее:\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n### Виндсерфинг\n\nДобавьте следующее в файл конфигурации Windsurf MCP.\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\nБолее подробную информацию см. [в документации Windsurf MCP](https://docs.windsurf.com/windsurf/mcp) .\n\n### Амазонка Бедрок\n\nВскоре....\n\n### Использование Docker вместо двоичного кода\n\nВскоре...\n\n## Использование\n\nПосле [установки](#installation) и [настройки](#setup-in-clients) сервера MCP вы можете начать использовать его с вашим любимым клиентом MCP.\n\nДиалог с ИИ-помощником можно начать с фразы:\n\n```\nUse MCP VictoriaMetrics in the following answers\n```\n\nНо это не обязательно, вы можете просто начать задавать вопросы, и помощник автоматически будет использовать инструменты и документацию, чтобы предоставить вам лучшие ответы. Просто взгляните на раздел [Пример диалога,](#dialog-example) чтобы лучше понять, что вы можете с ним сделать.\n\n### Набор инструментов\n\nMCP VictoriaMetrics предоставляет многочисленные инструменты для взаимодействия с вашим экземпляром VictoriaMetrics.\n\nВот список доступных инструментов:\n\n| Инструмент | Описание |\n| ---------------------------- | ------------------------------------------------------------------------- |\n| `query` | Выполнение мгновенных запросов PromQL/MetricsQL |\n| `query_range` | Выполнение ряда запросов PromQL/MetricsQL за определенный период времени |\n| `metrics` | Список доступных метрик |\n| `labels` | Список доступных названий меток |\n| `label_values` | Список значений для определенной метки |\n| `series` | Список доступных временных рядов |\n| `export` | Экспорт необработанных данных временного ряда в JSON или CSV |\n| `rules` | Просмотр правил оповещения и записи |\n| `alerts` | Просмотр текущих оповещений (активированных и ожидающих обработки) |\n| `flags` | Просмотр нестандартных флагов экземпляра VictoriaMetrics |\n| `metric_statistics` | Получить статистику использования метрик (в запросах) |\n| `active_queries` | Просмотреть текущие выполняемые запросы |\n| `top_queries` | Просмотреть наиболее частые или самые медленные запросы |\n| `tsdb_status` | Просмотр статистики мощности TSDB |\n| `tenants` | Список доступных арендаторов в многопользовательской кластерной настройке |\n| `documentation` | Поиск во встроенной документации VictoriaMetrics |\n| `metric_relabel_debug` | Отладка правил перемаркировки, совместимых с Prometheus |\n| `downsampling_filters_debug` | Отладка конфигурации понижения частоты дискретизации |\n| `retention_filters_debug` | Отладка конфигурации фильтров хранения |\n\n### Подсказки\n\nСервер включает в себя предопределенные подсказки для выполнения распространенных задач.\n\nНа данный момент это всего лишь примеры, библиотека подсказок будет дополнена в будущем:\n\n| Быстрый | Описание |\n| ------------------------------------------- | -------------------------------------------------------------------------------------------------------------------- |\n| `unused_metrics` | Найдите метрики, которые никогда не запрашиваются, и создайте конфигурации перемаркировки, чтобы прекратить их прием |\n| `documentation` | Поиск документации VictoriaMetrics по конкретным темам |\n| `rarely_used_metrics_with_high_cardinality` | Определите метрики с высокой кардинальностью, которые редко запрашиваются |\n\n## Пример диалога\n\nЭто всего лишь пример диалога с Assistant, использующим MCP Server, который может помочь понять, как он работает (но это реальный диалог с использованием MCP-сервера Victoria Metrics с Claude Desktop и настроенным на [общедоступный экземпляр игровой площадки](https://play.victoriametrics.com) ):\n\n* > **Вы** : Какие арендаторы доступны в моей виртуальной машине?\n* > **Помощник** *В вашем экземпляре VictoriaMetrics доступны следующие арендаторы `0:0` (арендатор по умолчанию), `1:0` , `10:0` , `100:0`*\n* > **Вы** : С этого момента, пожалуйста, используйте только арендатора по умолчанию.\n* > **Assistant** : *Я понимаю. С этого момента я буду использовать только арендатора по умолчанию (0:0) для всех операций с вашим экземпляром VictoriaMetrics.*\n* > **Вы** : Какие показатели у нас есть?\n* > **Assistant** : *В вашем экземпляре VictoriaMetrics доступно множество метрик. Список включает метрики из различных систем и компонентов, таких как:*\n >\n > * *Системные метрики ( `node_` , `machine_` и т.д.),*\n > * *Метрики Kubernetes ( `kube_` , `container_` , `apiserver_` и т. д.),*\n > * *Метрики, специфичные для VictoriaMetrics ( `vm_` , `vmagent_` , `vmalert_` и т. д.),*\n > * *Метрики времени выполнения Go ( `go_` ),*\n > * *Метрики, связанные с Prometheus ( `scrape_` , `promhttp_` ),*\n > * *Различные метрики приложения ( `argocd_` , `alertmanager_` ).*\n >\n > *Это полный список из более чем 2000 метрик, доступных для запроса. Есть ли конкретный тип метрик, который вы хотели бы изучить подробнее?*\n* > **Вы** : Хорошо, давайте работать с метриками kubernetes с этого момента. Пожалуйста, покажите мне, какие из этих метрик никогда не запрашивались?\n* > **Помощник** : *Основываясь на результатах, я вижу, что есть две метрики Kubernetes, которые никогда не запрашивались (количество запросов равно 0):*\n >\n > * *`kube_persistentvolumeclaim_deletion_timestamp` — эта метрика предоставляет информацию о том, когда PersistentVolumeClaims были помечены для удаления.*\n > * *`kube_pod_init_container_status_last_terminated_reason` — эта метрика показывает причину последнего завершения работы init-контейнеров в модулях.*\n >\n > *Эти показатели могут быть полезны для мониторинга и устранения неполадок:*\n >\n > * *Метрика временной метки удаления PVC может помочь отслеживать закономерности удаления или выявлять застрявшие PVC.*\n > * *Метрика причин завершения работы контейнера init может помочь выявить распространенные модели сбоев в контейнерах инициализации.*\n >\n > *Хотите, чтобы я подробнее объяснил, что отслеживают эти показатели, или предложил несколько полезных запросов, чтобы начать их мониторинг?*\n* > **Вы** : Да, пожалуйста. А также, можете ли вы создать правила перемаркировки, чтобы прекратить прием этих метрик?\n* > **Assistant** : *Я создам конфигурацию перемаркировки, чтобы исключить эти никогда не запрашиваемые метрики из сбора. Это поможет сократить накладные расходы на хранение и обработку. Вот конфигурация перемаркировки, которую вы можете добавить в конфигурацию скрапа VictoriaMetrics:*\n >\n > ```yaml\n > relabel_configs:\n > - action: drop\n > source_labels: [__name__]\n > regex: kube_persistentvolumeclaim_deletion_timestamp|kube_pod_init_container_status_last_terminated_reason\n > ```\n >\n > *Эта конфигурация будет:*\n >\n > 1. *Посмотрите на название каждой метрики (метка `__name__` )*\n > 2. *Проверьте, соответствует ли он хотя бы одной из указанных метрик.*\n > 3. *Удалить соответствующие показатели, не дав им собраться и сохраниться.*\n >\n > *Вы можете добавить это в файл конфигурации VictoriaMetrics или в соответствующие конфигурации заданий по очистке в конфигурации Prometheus, если вы используете Prometheus в качестве очистителя для VictoriaMetrics.* *Если вы хотите протестировать эту конфигурацию перемаркировки перед ее реализацией, вы можете использовать инструмент отладки перемаркировки метрики, чтобы увидеть, как она повлияет на ваши метрики.*\n* > **Вы** : Да, пожалуйста, отладьте эту конфигурацию перемаркировки в реальном времени, чтобы увидеть, как это повлияет на мои метрики.\n* > **Assistant** : *Отлично! Конфигурация перемаркировки работает так, как и ожидалось:*\n >\n > * *Первые два тестовых случая показывают, что метрики `kube_persistentvolumeclaim_deletion_timestamp` и `kube_pod_init_container_status_last_terminated_reason` корректно удаляются (результирующие метки пустые `{}` ).*\n > * *Третий тестовый случай подтверждает, что другие метрики, такие как `kube_pod_status_phase` , сохраняются и не затрагиваются нашим правилом удаления.*\n >\n > *Вы можете уверенно добавить эту конфигурацию перемаркировки в конфигурацию VictoriaMetrics или Prometheus, чтобы прекратить сбор этих неиспользуемых метрик. Это сэкономит место на диске и сократит накладные расходы на обработку, не влияя на ваши текущие возможности мониторинга.*\n* > **Вы** : Но я использую ресурс vmagent оператора VictoriaMetrics. Не могли бы вы предоставить мне часть спецификации для моего агента с этой конфигурацией переименования?\n* > **Assistant** : *Вы правы, поскольку вы используете оператор VictoriaMetrics и пользовательский ресурс VMAgent, конфигурация должна быть в соответствующем формате для этого CR. Вот как вы можете добавить конфигурацию relabel в свою спецификацию VMAgent:*\n >\n > ```yaml\n > apiVersion: operator.victoriametrics.com/v1beta1\n > kind: VMAgent\n > metadata:\n > name: your-vmagent-name\n > spec:\n > # ... your existing configuration ...\n > \n > # Add this section to configure global metric relabeling\n > inlineRelabelConfig:\n > - action: drop\n > sourceLabels: [\"__name__\"]\n > regex: kube_persistentvolumeclaim_deletion_timestamp|kube_pod_init_container_status_last_terminated_reason\n > \n > # ... rest of your configuration ...\n > ```\n >\n > *Ключевые моменты:*\n >\n > * *В спецификации VMAgent вы используете `inlineRelabelConfig` для правил встроенной перемаркировки.*\n > * *Поле называется `sourceLabels` (а не `source_labels` как в конфигурации Prometheus)*\n > * *Это глобальная перемаркировка, которая будет применена ко всем метрикам перед их сохранением.*\n* > **Вы** : Я решил не удалять эти метрики, а просто понизить частоту выборки для них с помощью Downsampling. Какие параметры downsampling у меня сейчас?\n* > **Помощник** : *Я вижу, что в настоящее время у вас следующая конфигурация понижения частоты дискретизации:*\n >\n > ```\n > -downsampling.period=\"7d:1m,30d:5m,180d:30m\"\n > ```\n >\n > *Это означает:*\n >\n > * *Данные старше 7 дней сохраняются в исходном разрешении.*\n > * *Данные возрастом от 7 до 30 дней сжимаются до разрешения в 1 минуту.*\n > * *Данные возрастом от 30 до 180 дней сжимаются до разрешения в 5 минут.*\n > * *Данные старше 180 дней сжимаются до 30-минутного разрешения.*\n >\n > *Теперь, если вы хотите добавить определенные правила понижения выборки для этих редко запрашиваемых метрик, вы можете использовать отладочный инструмент фильтров понижения выборки, чтобы проверить, как будут работать новые правила. Давайте протестируем конфигурацию, в которой мы применяем более агрессивную понижение выборки к этим конкретным метрикам:*\n >\n > *...получение серий и использование отладчика фильтра понижения частоты дискретизации...*\n >\n > *Отлично! Теперь наш тест работает правильно. На основе текущей конфигурации и нашего тестирования предлагаем следующую конфигурацию понижения частоты дискретизации для вашего VMAgent CR, которая применит более агрессивную понижение частоты дискретизации к редко запрашиваемым метрикам:*\n >\n > ```yaml\n > apiVersion: operator.victoriametrics.com/v1beta1\n > kind: VMCluster\n > metadata:\n > name: monitoring-vm-cluster\n > spec:\n > # ... your existing configuration ...\n > \n > # Update the vmstorage section with the new downsampling parameters\n > vmstorage:\n > extraArgs:\n > downsampling.period: '{__name__=\\\"kube_persistentvolumeclaim_deletion_timestamp\\\"}:1d:10m,{__name__=\\\"kube_pod_init_container_status_last_terminated_reason\\\"}:1d:10m,7d:1m,30d:5m,180d:30m'\n > \n > # ... rest of your configuration ...\n > ```\n >\n > *Эта конфигурация:*\n >\n > *- Применяет понижение разрешения на 10 минут для редко запрашиваемых метрик всего через 1 день.* *- Сохраняет существующую политику понижения разрешения для всех остальных метрик.*\n >\n > *Порядок правил понижения разрешения важен — более конкретные правила следует перечислить первыми.*\n\nВо время этого диалога помощник использовал следующие инструменты:\n\n* `tenants` , чтобы получить список доступных арендаторов\n* `documentation` для получения информации о функциональности и форматах данных\n* `metrics` для получения списка доступных метрик\n* `metrics_staistics` для получения информации об использовании метрик\n* `series` для получения временных рядов для отладки\n* `metric_relabel_debug` для отладки правил перемаркировки\n* `flags` для получения информации о параметрах экземпляра\n* `downsampling_filters_debug` для отладки конфигурации понижения частоты дискретизации\n\nНо вы можете использовать любые другие инструменты и комбинировать их по-своему.\n\n## Отказ от ответственности\n\nСлужбы и агенты ИИ вместе с серверами MCP, как этот, не могут гарантировать точность, полноту и надежность результатов. Вам следует дважды проверить результаты, полученные с помощью ИИ. Качество сервера MCP и его ответов во многом зависит от возможностей вашего клиента и качества модели, которую вы используете.\n\n## Внося вклад\n\nПриветствуются вклады в проект MCP VictoriaMetrics! Не стесняйтесь отправлять вопросы, запросы на функции или запросы на извлечение.","Russian","ru-RU",{"_27":64,"_31":66,"_35":67},{"_29":30,"_31":30,"_32":65},"# VictoriaMetrics MCP 服务器\n\n[VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/)的[模型上下文协议 (MCP)](https://modelcontextprotocol.io/)服务器的实现。\n\n这可以访问您的 VictoriaMetrics 实例，并与[VictoriaMetrics API](https://docs.victoriametrics.com/victoriametrics/url-examples/)和[文档](https://docs.victoriametrics.com/)无缝集成。它可以为您提供与 VictoriaMetrics 实例相关的监控、可观察性和调试任务的全面界面，并为工程师和工具提供高级自动化和交互功能。\n\n## 特征\n\n此 MCP 服务器允许您使用 VictoriaMetrics 的几乎所有只读 API，即[VMUI](https://docs.victoriametrics.com/#vmui)中可用的所有功能：\n\n* 查询指标和探索数据（如果客户端支持，甚至可以绘制图表）\n* 列出并导出可用的指标、标签、标签值和整个系列\n* 分析您的警报和记录规则和警报\n* 显示您的 VictoriaMetrics 实例的参数\n* 探索数据基数和指标使用情况统计\n* 分析您的查询\n* 调试重新标记规则、下采样和保留策略配置\n\n此外，MCP 服务器包含嵌入式最新文档，无需在线访问即可搜索。\n\n有关确切可用工具和提示的更多详细信息，请参阅[使用](#usage)部分。\n\n您可以将工具和文档搜索的功能组合到您的提示中，为您的 VictoriaMetrics 实例创建出色的使用场景。只需查看[对话框示例](#dialog-example)部分即可了解其工作原理。请注意，MCP 服务器及其响应的质量在很大程度上取决于您的客户端功能以及您所使用的模型的质量。\n\n您还可以将 MCP 服务器与其他可观察性或文档搜索相关的 MCP 服务器相结合，以获得更强大的结果。\n\n## 要求\n\n* [VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/)实例（[单节点](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/)或[集群](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/)）\n* 升级到 1.24 或更高版本（如果您想从源代码构建）\n\n## 安装\n\n### 去\n\n```bash\ngo install github.com/VictoriaMetrics-Community/mcp-victoriametrics/cmd/mcp-victoriametrics@latest\n```\n\n### 源代码\n\n```bash\ngit clone https://github.com/VictoriaMetrics-Community/mcp-victoriametrics.git\ncd mcp-victoriametrics\ngo build -o bin/mcp-victoriametrics ./cmd/mcp-victoriametrics/main.go\n\n# after that add bin/mcp-victoriametrics file to your PATH\n```\n\n### 二进制文件\n\n只需从[发布](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases)页面下载最新版本并将其放入您的 PATH 即可。\n\n### Docker\n\n即将推出...\n\n## 配置\n\nVictoriaMetrics 的 MCP 服务器通过环境变量进行配置：\n\n| 多变的 | 描述 | 必需的 | 默认 | 允许值 |\n| -------------------------- | --------------------------- | --- | ---------------- | -------------------- |\n| `VM_INSTANCE_ENTRYPOINT` | VictoriaMetrics 实例的 URL | 是的 | - | - |\n| `VM_INSTANCE_TYPE` | VictoriaMetrics 实例的类型 | 是的 | - | `single` 、 `cluster` |\n| `VM_INSTANCE_BEARER_TOKEN` | VictoriaMetrics API 的身份验证令牌 | 不 | - | - |\n| `MCP_SERVER_MODE` | 服务器运行模式 | 不 | `stdio` | `stdio` `sse` |\n| `MCP_SSE_ADDR` | SSE 服务器监听的地址 | 不 | `localhost:8080` | - |\n\n### 配置示例\n\n```bash\n# For a single-node instance\nexport VM_INSTANCE_ENTRYPOINT=\"http://localhost:8428\"\nexport VM_INSTANCE_TYPE=\"single\"\nexport VM_INSTANCE_BEARER_TOKEN=\"your-token\"\n\n# For a cluster\nexport VM_INSTANCE_ENTRYPOINT=\"https://play.victoriametrics.com\"\nexport VM_INSTANCE_TYPE=\"cluster\"\n\n# Server mode\nexport MCP_SERVER_MODE=\"sse\"\nexport MCP_SSE_ADDR=\"0.0.0.0:8080\"\n```\n\n## 在客户端中设置\n\n### 光标\n\n前往： `Settings` -> `Cursor Settings` -> `MCP` -> `Add new global MCP server` ，并将以下配置粘贴到您的 Cursor `~/.cursor/mcp.json`文件中：\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n请参阅[Cursor MCP 文档](https://docs.cursor.com/context/model-context-protocol)以了解更多信息。\n\n### 克劳德桌面\n\n将其添加到您的 Claude Desktop `claude_desktop_config.json`文件中（打开`Settings` -> `Developer` -> `Edit config`即可找到它）：\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n请参阅[Claude Desktop MCP 文档](https://modelcontextprotocol.io/quickstart/user)以了解更多信息。\n\n### 克劳德·科德\n\n运行以下命令：\n\n```sh\nclaude mcp add victoriametrics -- /path/to/mcp-victoriametrics \\\n -e VM_INSTANCE_ENTRYPOINT= \\\n -e VM_INSTANCE_TYPE=\n -e VM_INSTANCE_BEARER_TOKEN=\n```\n\n请参阅[Claude Code MCP 文档](https://docs.anthropic.com/en/docs/agents-and-tools/claude-code/tutorials#set-up-model-context-protocol-mcp)以了解更多信息。\n\n### Visual Studio 代码\n\n将其添加到您的 VS Code MCP 配置文件中：\n\n```json\n{\n \"servers\": {\n \"victoriametrics\": {\n \"type\": \"stdio\",\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n有关详细信息，请参阅[VS Code MCP 文档](https://code.visualstudio.com/docs/copilot/chat/mcp-servers)。\n\n### 泽德\n\n将以下内容添加到您的 Zed 配置文件：\n\n```json\n \"context_servers\": {\n \"victoriametrics\": {\n \"command\": {\n \"path\": \"/path/to/mcp-victoriametrics\",\n \"args\": [],\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n },\n \"settings\": {}\n }\n }\n```\n\n请参阅[Zed MCP 文档](https://zed.dev/docs/ai/mcp)以了解更多信息。\n\n### JetBrains IDE\n\n* 打开`Settings` -> `Tools` -> `AI Assistant` -> `Model Context Protocol (MCP)` 。\n* 单击`Add (+)`\n* 选择`As JSON`\n* 将以下内容输入到输入字段中：\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n### 风帆冲浪\n\n将以下内容添加到您的 Windsurf MCP 配置文件。\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\"\n }\n }\n }\n}\n```\n\n请参阅[Windsurf MCP 文档](https://docs.windsurf.com/windsurf/mcp)以了解更多信息。\n\n### 亚马逊基岩\n\n即将推出....\n\n### 使用 Docker 代替二进制文件\n\n即将推出...\n\n## 用法\n\n[安装](#installation)和[配置](#setup-in-clients)MCP 服务器后，您可以开始与您最喜欢的 MCP 客户端一起使用它。\n\n您可以从以下短语开始与AI助手对话：\n\n```\nUse MCP VictoriaMetrics in the following answers\n```\n\n但这不是必需的，您可以直接提问，助手会自动使用工具和文档为您提供最佳答案。您可以查看[对话框示例](#dialog-example)部分，更好地了解它的用途。\n\n### 工具集\n\nMCP VictoriaMetrics 提供了许多与您的 VictoriaMetrics 实例交互的工具。\n\n以下是可用工具的列表：\n\n| 工具 | 描述 |\n| ---------------------------- | ------------------------------ |\n| `query` | 执行即时 PromQL/MetricsQL 查询 |\n| `query_range` | 在一段时间内执行范围 PromQL/MetricsQL 查询 |\n| `metrics` | 列出可用指标 |\n| `labels` | 列出可用的标签名称 |\n| `label_values` | 列出特定标签的值 |\n| `series` | 列出可用的时间序列 |\n| `export` | 将原始时间序列数据导出为 JSON 或 CSV |\n| `rules` | 查看警报和记录规则 |\n| `alerts` | 查看当前警报（触发和待处理） |\n| `flags` | 查看 VictoriaMetrics 实例的非默认标志 |\n| `metric_statistics` | 获取指标使用情况（在查询中）统计信息 |\n| `active_queries` | 查看当前正在执行的查询 |\n| `top_queries` | 查看最频繁或最慢的查询 |\n| `tsdb_status` | 查看 TSDB 基数统计信息 |\n| `tenants` | 列出多租户集群设置中的可用租户 |\n| `documentation` | 在嵌入式 VictoriaMetrics 文档中搜索 |\n| `metric_relabel_debug` | 调试与 Prometheus 兼容的重新标记规则 |\n| `downsampling_filters_debug` | 调试下采样配置 |\n| `retention_filters_debug` | 调试保留过滤器配置 |\n\n### 提示\n\n该服务器包括常见任务的预定义提示。\n\n这些只是目前的示例，将来会添加到提示库中：\n\n| 迅速的 | 描述 |\n| ------------------------------------------- | --------------------------- |\n| `unused_metrics` | 查找从未查询过的指标并创建重新标记配置以停止提取它们 |\n| `documentation` | 在 VictoriaMetrics 文档中搜索特定主题 |\n| `rarely_used_metrics_with_high_cardinality` | 识别很少查询的高基数指标 |\n\n## 对话框示例\n\n这只是使用 MCP 服务器与 Assistant 对话的示例，有助于理解其工作原理（但这是使用带有 Claude Desktop 的 Victoria Metrics MCP 服务器并配置为[公共游乐场实例的](https://play.victoriametrics.com)真实对话）：\n\n* > **您**：我的虚拟机中有哪些可用的租户？\n* > **助手** *在您的 VictoriaMetrics 实例中，以下租户可用`0:0` （默认租户）、 `1:0` `10:0` `100:0`*\n* > **您**：从现在开始，请仅使用默认租户\n* > **助理**：*我明白了。从现在开始，我将只使用默认租户 (0:0) 来操作您的 VictoriaMetrics 实例。*\n* > **您**：我们有哪些指标？\n* > **助手**：*您的 VictoriaMetrics 实例中有许多可用的指标。该列表包含来自各种系统和组件的指标，例如：*\n >\n > * *系统指标（ `node_` 、 `machine_`等），*\n > * *Kubernetes 指标（ `kube_` 、 `container_` 、 `apiserver_`等），*\n > * *VictoriaMetrics 特定指标（ `vm_` 、 `vmagent_` 、 `vmalert_`等），*\n > * *Go 运行时指标（ `go_` ），*\n > * *Prometheus 相关指标（ `scrape_` 、 `promhttp_` ），*\n > * *各种应用程序指标（ `argocd_` 、 `alertmanager_` ）。*\n >\n > *这是一份包含 2000 多个可供查询的指标的完整列表。您是否有兴趣进一步了解某个特定类型的指标？*\n* > **你**：好的，我们现在开始使用 Kubernetes 指标。请告诉我哪些指标从未被查询过？\n* > **助手**：*根据结果，我可以看到有两个 Kubernetes 指标从未被查询过（查询计数为 0）：*\n >\n > * *`kube_persistentvolumeclaim_deletion_timestamp` - 此指标提供有关何时将 PersistentVolumeClaims 标记为删除的信息。*\n > * *`kube_pod_init_container_status_last_terminated_reason` - 此指标显示 pod 中的 init 容器最后终止的原因。*\n >\n > *这些指标对于监控和故障排除可能很有价值：*\n >\n > * *PVC 删除时间戳指标可以帮助跟踪删除模式或识别卡住的 PVC。*\n > * *初始化容器终止原因指标可以帮助识别初始化容器中的常见故障模式。*\n >\n > *您是否希望我进一步解释这些指标所跟踪的内容，或者建议一些有用的查询来开始监控它们？*\n* > **您**：好的，请问您能否创建重新标记规则来停止获取这些指标？\n* > **助手**：*我将创建一个重新标记配置，将这些从未查询过的指标从数据收集中排除。这将有助于减少存储和处理开销。以下是您可以添加到 VictoriaMetrics 抓取配置中的重新标记配置：*\n >\n > ```yaml\n > relabel_configs:\n > - action: drop\n > source_labels: [__name__]\n > regex: kube_persistentvolumeclaim_deletion_timestamp|kube_pod_init_container_status_last_terminated_reason\n > ```\n >\n > *此配置将：*\n >\n > 1. *查看每个指标的名称（ `__name__`标签）*\n > 2. *检查它是否符合指定的指标*\n > 3. *删除匹配的指标，防止它们被收集和存储*\n >\n > *您可以将其添加到 VictoriaMetrics 配置文件中，或者如果您使用 Prometheus 作为 VictoriaMetrics 的抓取工具，则将其添加到 Prometheus 配置中相关的抓取作业配置中。* *如果您想在实施此重新标记配置之前对其进行测试，可以使用指标重新标记调试工具来查看它将如何影响您的指标。*\n* > **您**：是的，请在实时序列上调试此重新标记配置，看看它将如何影响我的指标\n* > **助手**：*完美！relabel 配置按预期工作了：*\n >\n > * *前两个测试用例显示`kube_persistentvolumeclaim_deletion_timestamp`和`kube_pod_init_container_status_last_terminated_reason`指标均被正确删除（结果标签为空`{}` ）。*\n > * *第三个测试用例确认其他指标（如`kube_pod_status_phase`被保留并且不受我们的删除规则的影响。*\n >\n > *您可以放心地将此重新标记配置添加到您的 VictoriaMetrics 或 Prometheus 配置中，以停止收集这些未使用的指标。这将节省存储空间并降低处理开销，而不会影响您当前的监控功能。*\n* > **您**：但是我正在使用 VictoriaMetrics 操作员的 vmagent 资源，您能否为我提供使用此重新标记配置的代理的部分规范？\n* > **Assistant** ：*您说得对，由于您使用的是 VictoriaMetrics Operator 和 VMAgent 自定义资源，因此配置需要采用适合该 CR 的格式。以下是如何将 relabel 配置添加到 VMAgent 规范中：*\n >\n > ```yaml\n > apiVersion: operator.victoriametrics.com/v1beta1\n > kind: VMAgent\n > metadata:\n > name: your-vmagent-name\n > spec:\n > # ... your existing configuration ...\n > \n > # Add this section to configure global metric relabeling\n > inlineRelabelConfig:\n > - action: drop\n > sourceLabels: [\"__name__\"]\n > regex: kube_persistentvolumeclaim_deletion_timestamp|kube_pod_init_container_status_last_terminated_reason\n > \n > # ... rest of your configuration ...\n > ```\n >\n > *要点：*\n >\n > * *在 VMAgent 规范中，您可以使用`inlineRelabelConfig`来设置内联重新标记规则*\n > * *该字段称为`sourceLabels` （而不是 Prometheus 配置中的`source_labels` ）*\n > * *这是一个全局重新标记，将在存储之前应用于所有指标*\n* > **您**：我决定不删除这些指标，只是使用下采样来降低它们的采样率。我现在有哪些下采样参数？\n* > **助手**：*我可以看到您目前有以下下采样配置：*\n >\n > ```\n > -downsampling.period=\"7d:1m,30d:5m,180d:30m\"\n > ```\n >\n > *这意味着：*\n >\n > * *7 天以内的数据将保持其原始分辨率*\n > * *7-30 天之间的数据被下采样到 1 分钟的分辨率*\n > * *30-180 天之间的数据被下采样到 5 分钟分辨率*\n > * *超过 180 天的数据被下采样至 30 分钟分辨率*\n >\n > *现在，如果您想为那些很少查询的指标添加特定的下采样规则，可以使用下采样过滤器调试工具来测试新规则的运行情况。让我们测试一个配置，其中我们对这些特定指标应用更积极的下采样：*\n >\n > *...获取系列并使用下采样过滤器调试器...*\n >\n > *太棒了！现在我们的测试正常运行了。根据当前配置和我们的测试，我们为您的 VMAgent CR 提供了一个建议的下采样配置，该配置将对那些很少查询的指标应用更积极的下采样：*\n >\n > ```yaml\n > apiVersion: operator.victoriametrics.com/v1beta1\n > kind: VMCluster\n > metadata:\n > name: monitoring-vm-cluster\n > spec:\n > # ... your existing configuration ...\n > \n > # Update the vmstorage section with the new downsampling parameters\n > vmstorage:\n > extraArgs:\n > downsampling.period: '{__name__=\\\"kube_persistentvolumeclaim_deletion_timestamp\\\"}:1d:10m,{__name__=\\\"kube_pod_init_container_status_last_terminated_reason\\\"}:1d:10m,7d:1m,30d:5m,180d:30m'\n > \n > # ... rest of your configuration ...\n > ```\n >\n > *此配置：*\n >\n > *- 仅在 1 天后对很少查询的指标应用 10 分钟分辨率下采样* *- 对所有其他指标维持现有的下采样策略*\n >\n > *下采样规则的顺序很重要——应该首先列出更具体的规则。*\n\n在此对话中，助手使用了以下工具：\n\n* `tenants`获取可用租户列表\n* 获取有关功能和数据格式的信息的`documentation`\n* `metrics`获取可用指标列表\n* `metrics_staistics`获取有关指标使用情况的信息\n* `series`获取调试的时间序列\n* `metric_relabel_debug`用于调试重新标记规则\n* 获取实例参数信息的`flags`\n* `downsampling_filters_debug`用于调试下采样配置\n\n但您可以使用任何其他工具并以自己的方式组合它们。\n\n## 免责声明\n\n此类人工智能服务和代理以及 MCP 服务器无法保证结果的准确性、完整性和可靠性。您应该仔细检查使用人工智能获得的结果。MCP 服务器及其响应的质量很大程度上取决于客户端的功能以及您所使用的模型的质量。\n\n## 贡献\n\n欢迎为 MCP VictoriaMetrics 项目做出贡献！欢迎提交问题、功能请求或拉取请求。","Chinese","zh-CN","githubUser","locale",{"_71":72,"_13":73,"_80":30,"_81":30,"_16":82,"_83":84,"_31":30,"_110":111,"_118":-5,"_119":-5,"_120":121,"_183":184,"_189":-5,"_91":30,"_190":168,"_94":191,"_192":193},"addedAt","2025-05-11T20:53:34.947732Z",[74,75,76,77,78,79],"category:observability","category:monitoring","category:documentation-access","hosting:remote-capable","language:go","license_grade:a","descriptionMarkdown","descriptionPlainText",6578,"integrations",[85,96,103],{"_86":87,"_29":93,"_94":95},"brand",{"_88":89,"_31":90,"_91":92},"logoSvgPath","M10.204 14.35l.007.01-.999 2.413a5.171 5.171 0 0 1-2.075-2.597l2.578-.437.004.005a.44.44 0 0 1 .484.606zm-.833-2.129a.44.44 0 0 0 .173-.756l.002-.011L7.585 9.7a5.143 5.143 0 0 0-.73 3.255l2.514-.725.002-.009zm1.145-1.98a.44.44 0 0 0 .699-.337l.01-.005.15-2.62a5.144 5.144 0 0 0-3.01 1.442l2.147 1.523.004-.002zm.76 2.75l.723.349.722-.347.18-.78-.5-.623h-.804l-.5.623.179.779zm1.5-3.095a.44.44 0 0 0 .7.336l.008.003 2.134-1.513a5.188 5.188 0 0 0-2.992-1.442l.148 2.615.002.001zm10.876 5.97l-5.773 7.181a1.6 1.6 0 0 1-1.248.594l-9.261.003a1.6 1.6 0 0 1-1.247-.596l-5.776-7.18a1.583 1.583 0 0 1-.307-1.34L2.1 5.573c.108-.47.425-.864.863-1.073L11.305.513a1.606 1.606 0 0 1 1.385 0l8.345 3.985c.438.209.755.604.863 1.073l2.062 8.955c.108.47-.005.963-.308 1.34zm-3.289-2.057c-.042-.01-.103-.026-.145-.034-.174-.033-.315-.025-.479-.038-.35-.037-.638-.067-.895-.148-.105-.04-.18-.165-.216-.216l-.201-.059a6.45 6.45 0 0 0-.105-2.332 6.465 6.465 0 0 0-.936-2.163c.052-.047.15-.133.177-.159.008-.09.001-.183.094-.282.197-.185.444-.338.743-.522.142-.084.273-.137.415-.242.032-.024.076-.062.11-.089.24-.191.295-.52.123-.736-.172-.216-.506-.236-.745-.045-.034.027-.08.062-.111.088-.134.116-.217.23-.33.35-.246.25-.45.458-.673.609-.097.056-.239.037-.303.033l-.19.135a6.545 6.545 0 0 0-4.146-2.003l-.012-.223c-.065-.062-.143-.115-.163-.25-.022-.268.015-.557.057-.905.023-.163.061-.298.068-.475.001-.04-.001-.099-.001-.142 0-.306-.224-.555-.5-.555-.275 0-.499.249-.499.555l.001.014c0 .041-.002.092 0 .128.006.177.044.312.067.475.042.348.078.637.056.906a.545.545 0 0 1-.162.258l-.012.211a6.424 6.424 0 0 0-4.166 2.003 8.373 8.373 0 0 1-.18-.128c-.09.012-.18.04-.297-.029-.223-.15-.427-.358-.673-.608-.113-.12-.195-.234-.329-.349-.03-.026-.077-.062-.111-.088a.594.594 0 0 0-.348-.132.481.481 0 0 0-.398.176c-.172.216-.117.546.123.737l.007.005.104.083c.142.105.272.159.414.242.299.185.546.338.743.522.076.082.09.226.1.288l.16.143a6.462 6.462 0 0 0-1.02 4.506l-.208.06c-.055.072-.133.184-.215.217-.257.081-.546.11-.895.147-.164.014-.305.006-.48.039-.037.007-.09.02-.133.03l-.004.002-.007.002c-.295.071-.484.342-.423.608.061.267.349.429.645.365l.007-.001.01-.003.129-.029c.17-.046.294-.113.448-.172.33-.118.604-.217.87-.256.112-.009.23.069.288.101l.217-.037a6.5 6.5 0 0 0 2.88 3.596l-.09.218c.033.084.069.199.044.282-.097.252-.263.517-.452.813-.091.136-.185.242-.268.399-.02.037-.045.095-.064.134-.128.275-.034.591.213.71.248.12.556-.007.69-.282v-.002c.02-.039.046-.09.062-.127.07-.162.094-.301.144-.458.132-.332.205-.68.387-.897.05-.06.13-.082.215-.105l.113-.205a6.453 6.453 0 0 0 4.609.012l.106.192c.086.028.18.042.256.155.136.232.229.507.342.84.05.156.074.295.145.457.016.037.043.09.062.129.133.276.442.402.69.282.247-.118.341-.435.213-.71-.02-.039-.045-.096-.065-.134-.083-.156-.177-.261-.268-.398-.19-.296-.346-.541-.443-.793-.04-.13.007-.21.038-.294-.018-.022-.059-.144-.083-.202a6.499 6.499 0 0 0 2.88-3.622c.064.01.176.03.213.038.075-.05.144-.114.28-.104.266.039.54.138.87.256.154.06.277.128.448.173.036.01.088.019.13.028l.009.003.007.001c.297.064.584-.098.645-.365.06-.266-.128-.537-.423-.608zM16.4 9.701l-1.95 1.746v.005a.44.44 0 0 0 .173.757l.003.01 2.526.728a5.199 5.199 0 0 0-.108-1.674A5.208 5.208 0 0 0 16.4 9.7zm-4.013 5.325a.437.437 0 0 0-.404-.232.44.44 0 0 0-.372.233h-.002l-1.268 2.292a5.164 5.164 0 0 0 3.326.003l-1.27-2.296h-.01zm1.888-1.293a.44.44 0 0 0-.27.036.44.44 0 0 0-.214.572l-.003.004 1.01 2.438a5.15 5.15 0 0 0 2.081-2.615l-2.6-.44-.004.005z","Kubernetes","slug","kubernetes","Enables monitoring and analysis of Kubernetes metrics, including the ability to query container, pod, and other Kubernetes-specific metrics collected by VictoriaMetrics.","uid","c45o14g91q",{"_86":97,"_29":101,"_94":102},{"_88":98,"_31":99,"_91":100},"M12 0C5.373 0 0 5.372 0 12c0 6.627 5.373 12 12 12s12-5.373 12-12c0-6.628-5.373-12-12-12zm0 22.46c-1.885 0-3.414-1.26-3.414-2.814h6.828c0 1.553-1.528 2.813-3.414 2.813zm5.64-3.745H6.36v-2.046h11.28v2.046zm-.04-3.098H6.391c-.037-.043-.075-.086-.111-.13-1.155-1.401-1.427-2.133-1.69-2.879-.005-.025 1.4.287 2.395.511 0 0 .513.119 1.262.255-.72-.843-1.147-1.915-1.147-3.01 0-2.406 1.845-4.508 1.18-6.207.648.053 1.34 1.367 1.387 3.422.689-.951.977-2.69.977-3.755 0-1.103.727-2.385 1.454-2.429-.648 1.069.168 1.984.894 4.256.272.854.237 2.29.447 3.201.07-1.892.395-4.652 1.595-5.605-.529 1.2.079 2.702.494 3.424.671 1.164 1.078 2.047 1.078 3.716a4.642 4.642 0 01-1.11 2.996c.792-.149 1.34-.283 1.34-.283l2.573-.502s-.374 1.538-1.81 3.019z","Prometheus","prometheus","Supports Prometheus-compatible functionality including querying with PromQL, metric relabeling rules debugging, and integration with Prometheus configuration when used as a scraper for VictoriaMetrics.","abdjuydy4l",{"_86":104,"_29":108,"_94":109},{"_88":105,"_31":106,"_91":107},"M1.893 3.078C.239 1.537 4.846.009 11.963 0h.076c7.116.009 11.723 1.537 10.067 3.078 0 0-6.168 5.655-8.301 7.473-.404.346-1.084.607-1.766.616h-.076c-.682-.009-1.362-.271-1.767-.616-2.132-1.818-8.301-7.473-8.303-7.473Zm20.549 2.771V8.58c0 .3-.115.698-.319.885-1.332 1.222-6.47 5.925-8.319 7.502-.405.345-1.085.606-1.767.615h-.072c-.683-.009-1.362-.27-1.767-.615-1.849-1.577-6.987-6.28-8.32-7.502-.204-.187-.318-.585-.318-.885V5.849c1.96 1.788 7.163 6.505 8.638 7.764.404.346 1.084.607 1.767.616h.072c.682-.009 1.362-.271 1.767-.616 1.474-1.258 6.678-5.973 8.638-7.764Zm0 6.418v2.73c0 .301-.115.698-.319.885-1.332 1.222-6.47 5.926-8.319 7.502-.405.346-1.085.607-1.767.616h-.072c-.683-.009-1.362-.271-1.767-.616-1.849-1.576-6.987-6.28-8.32-7.502-.204-.187-.318-.585-.318-.885v-2.73c1.96 1.788 7.163 6.505 8.638 7.764.404.346 1.084.606 1.767.615h.072c.682-.009 1.362-.27 1.767-.615 1.474-1.258 6.678-5.976 8.638-7.764Z","VictoriaMetrics","victoriametrics","Provides access to a VictoriaMetrics monitoring system, enabling querying metrics, exploring data, analyzing alerts/rules, viewing instance parameters, exploring data cardinality, analyzing queries, and debugging relabeling rules and retention configurations.","o9aqcpua08","namespace",{"_112":113,"_91":117},"owner",{"_114":115},"githubAccount",{"_116":117},"login","VictoriaMetrics-Community","readmeMarkdown","redditPostExternalId","repository",{"_122":123,"_149":150,"_154":155,"_174":-5,"_175":176,"_177":-5,"_178":179},"argumentsJsonSchema",{"_124":125,"_146":147,"_131":148},"properties",{"_126":127,"_133":134,"_137":138,"_140":141,"_143":144},"MCP_SSE_ADDR",{"_128":129,"_29":130,"_131":132},"default","localhost:8080","Address for SSE server to listen on","type","string","MCP_SERVER_MODE",{"_128":135,"_29":136,"_131":132},"stdio","Server operation mode","VM_INSTANCE_TYPE",{"_29":139,"_131":132},"Type of VictoriaMetrics instance","VM_INSTANCE_ENTRYPOINT",{"_29":142,"_131":132},"URL to VictoriaMetrics instance","VM_INSTANCE_BEARER_TOKEN",{"_29":145,"_131":132},"Authentication token for VictoriaMetrics API","required",[140,137],"object","githubProject",{"_32":151,"_152":153},"# VictoriaMetrics MCP Server\n\n[![Latest Release](https://img.shields.io/github/v/release/VictoriaMetrics-Community/mcp-victoriametrics?sort=semver&label=&filter=!*-victorialogs&logo=github&labelColor=gray&color=gray&link=https%3A%2F%2Fgithub.com%2FVictoriaMetrics-Community%2Fmcp-victoriametrics%2Freleases%2Flatest)](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases)\n[![Trust Score](https://archestra.ai/mcp-catalog/api/badge/quality/VictoriaMetrics-Community/mcp-victoriametrics)](https://archestra.ai/mcp-catalog/victoriametrics-community__mcp-victoriametrics)\n[![smithery badge](https://smithery.ai/badge/@VictoriaMetrics-Community/mcp-victoriametrics)](https://smithery.ai/server/@VictoriaMetrics-Community/mcp-victoriametrics)\n![License](https://img.shields.io/github/license/VictoriaMetrics-Community/mcp-victoriametrics?labelColor=green&label=&link=https%3A%2F%2Fgithub.com%2FVictoriaMetrics-Community%2Fmcp-victoriametrics%2Fblob%2Fmain%2FLICENSE)\n![Slack](https://img.shields.io/badge/Join-4A154B?logo=slack&link=https%3A%2F%2Fslack.victoriametrics.com)\n![X](https://img.shields.io/twitter/follow/VictoriaMetrics?style=flat&label=Follow&color=black&logo=x&labelColor=black&link=https%3A%2F%2Fx.com%2FVictoriaMetrics)\n![Reddit](https://img.shields.io/reddit/subreddit-subscribers/VictoriaMetrics?style=flat&label=Join&labelColor=red&logoColor=white&logo=reddit&link=https%3A%2F%2Fwww.reddit.com%2Fr%2FVictoriaMetrics)\n\nThe implementation of [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) server for [VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/).\n\nThis provides access to your VictoriaMetrics instance and seamless integration with [VictoriaMetrics APIs](https://docs.victoriametrics.com/victoriametrics/url-examples/) and [documentation](https://docs.victoriametrics.com/).\nIt can give you a comprehensive interface for monitoring, observability, and debugging tasks related to your VictoriaMetrics instances, enable advanced automation and interaction capabilities for engineers and tools.\n\n## Features\n\nThis MCP server allows you to use almost all read-only APIs of VictoriaMetrics, i.e. all functions available in [VMUI](https://docs.victoriametrics.com/#vmui):\n\n- Querying metrics and exploring data (even drawing graphs if your client supports it)\n- Listing and exporting available metrics, labels, labels values and entire series\n- Analyzing and testing your alerting and recording rules and alerts\n- Showing parameters of your VictoriaMetrics instance\n- Exploring cardinality of your data and metrics usage statistics\n- Analyzing, tracing, prettifying and explaining your queries\n- Debugging your relabeling rules, downsampling and retention policy configurations \n- Integration with [VictoriaMetrics Cloud](https://docs.victoriametrics.com/victoriametrics-cloud/)\n \nIn addition, the MCP server contains embedded up-to-date documentation and is able to search it without online access.\n\nMore details about the exact available tools and prompts can be found in the [Usage](#usage) section.\n\nYou can combine functionality of tools, docs search in your prompts and invent great usage scenarios for your VictoriaMetrics instance.\nJust check the [Dialog example](#dialog-example) section to see how it can work.\nAnd please note the fact that the quality of the MCP Server and its responses depends very much on the capabilities of your client and the quality of the model you are using.\n\nYou can also combine the MCP server with other observability or doc search related MCP Servers and get even more powerful results.\n\n## Try without installation\n\nThere is a publicly available instance of the VictoriaMetrics MCP Server that you can use to test the features without installing it: \n\n```\nhttps://play-mcp.victoriametrics.com/mcp\n```\n\n**Attention!** This URL is not supposed to be opened in a browser, it is intended to be used in MCP clients.\n\nIt's available in [Streamable HTTP](#modes) mode and configured to work with [Public VictoriaMetrics Playground](https://play.victoriametrics.com).\n\nHere is example of configuration for [Claude Desktop](https://claude.ai/download):\n\n![image](https://github.com/user-attachments/assets/b9cca3ff-f4c3-47bc-aac7-9359f12e858e)\n\n## Requirements\n\n- [VictoriaMetrics](https://docs.victoriametrics.com/victoriametrics/) or [VictoriaMetrics Cloud](https://docs.victoriametrics.com/victoriametrics-cloud/) instance ([single-node](https://docs.victoriametrics.com/victoriametrics/single-server-victoriametrics/) or [cluster](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/))\n- Go 1.24 or higher (if you want to build from source)\n\n## Installation\n\n### Go\n\n```bash\ngo install github.com/VictoriaMetrics-Community/mcp-victoriametrics/cmd/mcp-victoriametrics@latest\n```\n\n### Binaries\n\nJust download the latest release from [Releases](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases) page and put it to your PATH.\n\nExample for Linux x86_64 (note that other architectures and platforms are also available):\n\n```bash\nlatest=$(curl -s https://api.github.com/repos/VictoriaMetrics-Community/mcp-victoriametrics/releases/latest | grep 'tag_name' | cut -d\\\" -f4)\nwget https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases/download/$latest/mcp-victoriametrics_Linux_x86_64.tar.gz\ntar axvf mcp-victoriametrics_Linux_x86_64.tar.gz\n```\n\n### Docker\n\nYou can run VictoriaMetrics MCP Server using Docker. \n\nThis is the easiest way to get started without needing to install Go or build from source.\n\n```bash\ndocker run -d --name mcp-victoriametrics \\\n -e MCP_SERVER_MODE=sse \\\n -e VM_INSTANCE_ENTRYPOINT=https://play.victoriametrics.com \\\n -e VM_INSTANCE_TYPE=cluster \\\n ghcr.io/victoriametrics-community/mcp-victoriametrics\n```\n\nYou should replace environment variables with your own parameters.\n\nNote that the `MCP_SERVER_MODE=sse` flag is used to enable Server-Sent Events mode, which used by MCP clients to connect.\nAlternatively, you can use `MCP_SERVER_MODE=http` to enable Streamable HTTP mode. More details about server modes can be found in the [Configuration](#configuration) section.\n\nSee available docker images in [github registry](https://github.com/orgs/VictoriaMetrics-Community/packages/container/package/mcp-victoriametrics).\n\nAlso see [Using Docker instead of binary](#using-docker-instead-of-binary) section for more details about using Docker with MCP server with clients in stdio mode.\n\n### Source Code\n\nFor building binary from source code you can use the following approach:\n\n- Clone repo:\n \n ```bash\n git clone https://github.com/VictoriaMetrics-Community/mcp-victoriametrics.git\n cd mcp-victoriametrics\n ```\n- Build binary from cloned source code: \n \n ```bash\n make build\n # after that you can find binary mcp-victoriametrics and copy this file to your PATH or run inplace\n ```\n- Build image from cloned source code:\n \n ```bash\n docker build -t mcp-victoriametrics .\n # after that you can use docker image mcp-victoriametrics for running or pushing\n ```\n\n### Smithery\n\nTo install VictoriaMetrics MCP Server for your client automatically via Smithery, yo can use the following commands:\n\n```bash\n# Get the list of supported MCP clients\nnpx -y @smithery/cli list clients\n#Available clients:\n# claude\n# cline\n# windsurf\n# roocode\n# witsy\n# enconvo\n# cursor\n# vscode\n# vscode-insiders\n# boltai\n# amazon-bedrock\n\n# Install VictoriaMetrics MCP server for your client\nnpx -y @smithery/cli install @VictoriaMetrics-Community/mcp-victoriametrics --client \n# and follow the instructions\n```\n\n## Configuration\n\nMCP Server for VictoriaMetrics is configured via environment variables:\n\n| Variable | Description | Required | Default | Allowed values |\n|------------------------------------------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|----------------------------------------|------------------|------------------------|\n| `VM_INSTANCE_ENTRYPOINT` / `VMC_API_KEY` | URL to VictoriaMetrics instance (it should be root `/` URL of vmsingle or vmselect) | Yes (if you don't use `VMC_API_KEY`) | - | - |\n| `VM_INSTANCE_TYPE` | Type of VictoriaMetrics instance | Yes (if you don't use ``VMC_API_KEY``) | - | `single`, `cluster` |\n| `VM_INSTANCE_BEARER_TOKEN` | Authentication token for VictoriaMetrics API | No | - | - |\n| `VM_INSTANCE_HEADERS` | Custom HTTP headers to send with requests (comma-separated key=value pairs) | No | - | - |\n| `VMC_API_KEY` | [API key from VictoriaMetrics Cloud Console](https://docs.victoriametrics.com/victoriametrics-cloud/api/) | No | - | - |\n| `MCP_SERVER_MODE` | Server operation mode. See [Modes](#modes) for details. | No | `stdio` | `stdio`, `sse`, `http` |\n| `MCP_LISTEN_ADDR` | Address for SSE or HTTP server to listen on | No | `localhost:8080` | - |\n| `MCP_DISABLED_TOOLS` | Comma-separated list of tools to disable | No | - | - |\n| `MCP_DISABLE_RESOURCES` | Disable all resources (documentation tool will continue to work) | No | `false` | `false`, `true` | \n| `MCP_HEARTBEAT_INTERVAL` | Defines the heartbeat interval for the streamable-http protocol.
It means the MCP server will send a heartbeat to the client through the GET connection,
to keep the connection alive from being closed by the network infrastructure (e.g. gateways) | No | `30s` | - |\n\nYou can use two options to connect to your VictoriaMetrics instance:\n\n- Using `VM_INSTANCE_ENTRYPOINT` + `VM_INSTANCE_TYPE` + `VM_INSTANCE_BEARER_TOKEN` (optional) environment variables to connect to any single-node or cluster instance of VictoriaMetrics.\n- Using `VMC_API_KEY` environment variable to work with your [VictoriaMetrics Cloud](https://victoriametrics.com/products/cloud/) instances.\n\n### Modes\n\nMCP Server supports the following modes of operation (transports):\n\n- `stdio` - Standard input/output mode, where the server reads commands from standard input and writes responses to standard output. This is the default mode and is suitable for local servers.\n- `sse` - Server-Sent Events. Server will expose the `/sse` and `/message` endpoints for SSE connections.\n- `http` - Streamable HTTP. Server will expose the `/mcp` endpoint for HTTP connections.\n\nMore info about traqnsports you can find in MCP docs:\n\n- [Core concepts -> Transports](https://modelcontextprotocol.io/docs/concepts/transports)\n- [Specifications -> Transports](https://modelcontextprotocol.io/specification/2025-03-26/basic/transports)\n\n### Сonfiguration examples\n\n```bash\n# For a single-node instance\nexport VM_INSTANCE_ENTRYPOINT=\"http://localhost:8428\"\nexport VM_INSTANCE_TYPE=\"single\"\nexport VM_INSTANCE_BEARER_TOKEN=\"your-token\"\n\n# For a cluster\nexport VM_INSTANCE_ENTRYPOINT=\"https://play.victoriametrics.com\"\nexport VM_INSTANCE_TYPE=\"cluster\"\nexport MCP_DISABLED_TOOLS=\"export,metric_statistics,test_rules\" # disable export, statistics and rules unit test tools\n\n# For VictoriaMetrics Cloud\nexport VMC_API_KEY=\"\"\n\n# Server mode\nexport MCP_SERVER_MODE=\"sse\"\nexport MCP_LISTEN_ADDR=\"0.0.0.0:8080\"\n\n# Custom headers for authentication (e.g., behind a reverse proxy)\n# Expected syntax is key=value separated by commas\nexport VM_INSTANCE_HEADERS=\"

=\"\n```\n\n## Endpoints\n\nIn SSE and HTTP modes the MCP server provides the following endpoints:\n\n| Endpoint | Description |\n|----------------------|---------------------------------------------------------------------------------------------------|\n| `/sse` + `/message` | Endpoints for messages in SSE mode (for MCP clients that support SSE) |\n| `/mcp` | HTTP endpoint for streaming messages in HTTP mode (for MCP clients that support Streamable HTTP) |\n| `/metrics` | Metrics in Prometheus format for monitoring the MCP server |\n| `/health/liveness` | Liveness check endpoint to ensure the server is running |\n| `/health/readiness` | Readiness check endpoint to ensure the server is ready to accept requests |\n\n## Setup in clients\n\n### Cursor\n\nGo to: `Settings` -> `Cursor Settings` -> `MCP` -> `Add new global MCP server` and paste the following configuration into your Cursor `~/.cursor/mcp.json` file:\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\",\n \"VM_INSTANCE_HEADERS\": \"

=\"\n }\n }\n }\n}\n```\n\nSee [Cursor MCP docs](https://docs.cursor.com/context/model-context-protocol) for more info.\n\n### Claude Desktop\n\nAdd this to your Claude Desktop `claude_desktop_config.json` file (you can find it if open `Settings` -> `Developer` -> `Edit config`):\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\",\n \"VM_INSTANCE_HEADERS\": \"

=\"\n }\n }\n }\n}\n```\n\nSee [Claude Desktop MCP docs](https://modelcontextprotocol.io/quickstart/user) for more info.\n\n### Claude Code\n\nRun the command:\n\n```sh\nclaude mcp add victoriametrics -- /path/to/mcp-victoriametrics \\\n -e VM_INSTANCE_ENTRYPOINT= \\\n -e VM_INSTANCE_TYPE=\n -e VM_INSTANCE_BEARER_TOKEN=\n -e VM_INSTANCE_HEADERS=\"

=\"\n```\n\nSee [Claude Code MCP docs](https://docs.anthropic.com/en/docs/agents-and-tools/claude-code/tutorials#set-up-model-context-protocol-mcp) for more info.\n\n### Visual Studio Code\n\nAdd this to your VS Code MCP config file:\n\n```json\n{\n \"servers\": {\n \"victoriametrics\": {\n \"type\": \"stdio\",\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\",\n \"VM_INSTANCE_HEADERS\": \"

=\"\n }\n }\n }\n}\n```\n\nSee [VS Code MCP docs](https://code.visualstudio.com/docs/copilot/chat/mcp-servers) for more info.\n\n### Zed\n\nAdd the following to your Zed config file:\n\n```json\n \"context_servers\": {\n \"victoriametrics\": {\n \"command\": {\n \"path\": \"/path/to/mcp-victoriametrics\",\n \"args\": [],\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\",\n \"VM_INSTANCE_HEADERS\": \"

=\"\n }\n },\n \"settings\": {}\n }\n }\n```\n\nSee [Zed MCP docs](https://zed.dev/docs/ai/mcp) for more info.\n\n### JetBrains IDEs\n\n- Open `Settings` -> `Tools` -> `AI Assistant` -> `Model Context Protocol (MCP)`.\n- Click `Add (+)`\n- Select `As JSON`\n- Put the following to the input field:\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\",\n \"VM_INSTANCE_HEADERS\": \"

=\"\n }\n }\n }\n}\n```\n\n### Windsurf\n\nAdd the following to your Windsurf MCP config file.\n\n```json\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"/path/to/mcp-victoriametrics\",\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\",\n \"VM_INSTANCE_HEADERS\": \"

=\"\n }\n }\n }\n}\n```\n\nSee [Windsurf MCP docs](https://docs.windsurf.com/windsurf/mcp) for more info.\n\n### Using Docker instead of binary\n\nYou can run VictoriaMetrics MCP server using Docker instead of local binary.\n\nYou should replace run command in configuration examples above in the following way:\n\n```\n{\n \"mcpServers\": {\n \"victoriametrics\": {\n \"command\": \"docker\",\n \"args\": [\n \"run\",\n \"-i\", \"--rm\",\n \"-e\", \"VM_INSTANCE_ENTRYPOINT\",\n \"-e\", \"VM_INSTANCE_TYPE\",\n \"-e\", \"VM_INSTANCE_BEARER_TOKEN\",\n \"-e\", \"VM_INSTANCE_HEADERS\",\n \"ghcr.io/victoriametrics-community/mcp-victoriametrics\",\n ],\n \"env\": {\n \"VM_INSTANCE_ENTRYPOINT\": \"\",\n \"VM_INSTANCE_TYPE\": \"\",\n \"VM_INSTANCE_BEARER_TOKEN\": \"\",\n \"VM_INSTANCE_HEADERS\": \"

=\"\n }\n }\n }\n}\n```\n\n## Usage\n\nAfter [installing](#installation) and [configuring](#setup-in-clients) the MCP server, you can start using it with your favorite MCP client.\n\nYou can start dialog with AI assistant from the phrase:\n\n```\nUse MCP VictoriaMetrics in the following answers\n```\n\nBut it's not required, you can just start asking questions and the assistant will automatically use the tools and documentation to provide you with the best answers.\nJust take a look into [Dialog example](#dialog-example) section for better understanding what you can do with it.\n\n### Toolset\n\nMCP VictoriaMetrics provides numerous tools for interacting with your VictoriaMetrics instance.\n\nHere's a list of common available tools:\n\n| Tool | Description |\n|------------------------------|-----------------------------------------------------------|\n| `query` | Execute instant PromQL/MetricsQL queries |\n| `query_range` | Execute range PromQL/MetricsQL queries over a time period |\n| `metrics` | List available metrics |\n| `labels` | List available label names |\n| `label_values` | List values for a specific label |\n| `series` | List available time series |\n| `export` | Export raw time series data to JSON or CSV |\n| `rules` | View alerting and recording rules |\n| `alerts` | View current alerts (firing and pending) |\n| `flags` | View non-default flags of the VictoriaMetrics instance |\n| `metric_statistics` | Get metrics usage (in queries) statistics |\n| `active_queries` | View currently executing queries |\n| `top_queries` | View most frequent or slowest queries |\n| `tsdb_status` | View TSDB cardinality statistics |\n| `tenants` | List available tenants in multi-tenant cluster setup |\n| `documentation` | Search in embedded VictoriaMetrics documentation |\n| `metric_relabel_debug` | Debug Prometheus-compatible relabeling rules |\n| `downsampling_filters_debug` | Debug downsampling configuration |\n| `retention_filters_debug` | Debug retention filters configuration |\n| `prettify_query` | Prettify and format PromQL/MetricsQL queries |\n| `explain_query` | Parse PromQL/MetricsQL queries and explain how it works |\n| `test_rules` | Unit-test alerting and recording rules using vmalert tool |\n\nHere are some additional tools that are available for [VictoriaMetrics Cloud](https://docs.victoriametrics.com/victoriametrics-cloud/) (with specifying `VMC_API_KEY` parameter) users:\n\n| Tool | Description |\n|-------------------|--------------------------------------------------------------------------------|\n| `deployments` | List available deployments in VictoriaMetrics Cloud |\n| `cloud_providers` | List available cloud providers in VictoriaMetrics Cloud |\n| `regions` | List available cloud providers regions in VictoriaMetrics Cloud |\n| `tiers` | List available deployment tiers in VictoriaMetrics Cloud |\n| `access_tokens` | List available deployment access tokens in VictoriaMetrics Cloud |\n| `rule_filenames` | List available alerting and recording rule filenames in VictoriaMetrics Cloud |\n| `rule_file` | Content of a specific alerting or recording rule file in VictoriaMetrics Cloud |\n\n### Prompts\n\nThe server includes pre-defined prompts for common tasks.\n\nThese are just examples at the moment, the prompt library will be added to in the future:\n\n| Prompt | Description |\n|--------|-------------|\n| `unused_metrics` | Find metrics that are never queried and create relabel configs to stop ingesting them |\n| `documentation` | Search VictoriaMetrics documentation for specific topics |\n| `rarely_used_metrics_with_high_cardinality` | Identify high-cardinality metrics that are rarely queried |\n\n## Dialog example\n\nThis is just example of dialog with Assistant using MCP Server that can help to understand how it can work (but's it's real dialog with using the Victoria Metrics MCP server with Claude Desktop and configured to [public playground instance](https://play.victoriametrics.com)):\n\n- > **You**: What tenants are available in my VM?\n- > **Assistant** *In your VictoriaMetrics instance, the following tenants are available `0:0` (default tenant), `1:0`, `10:0`, `100:0`*\n- > **You**: From now on, please use only default tenant\n- > **Assistant**: *I understand. From now on, I'll use only the default tenant (0:0) for all operations with your VictoriaMetrics instance.*\n- > **You**: Which metrics do we have?\n- > **Assistant**: *There are many metrics available in your VictoriaMetrics instance. The list includes metrics from various systems and components, such as:*\n >\n > - *System metrics (`node_`, `machine_`, etc.),*\n > - *Kubernetes metrics (`kube_`, `container_`, `apiserver_`, etc.),*\n > - *VictoriaMetrics-specific metrics (`vm_`, `vmagent_`, `vmalert_`, etc.),*\n > - *Go runtime metrics (`go_`),*\n > - *Prometheus-related metrics (`scrape_`, `promhttp_`),*\n > - *Various application metrics (`argocd_`, `alertmanager_`).*\n >\n > *This is a comprehensive list of over 2000 metrics that are available for querying. Is there a specific type of metric you're interested in exploring further?*\n- > **You**: Ok, let's work with kubernetes metrics from now on. Please show me which of these metrics never queried?\n- > **Assistant**: *Based on the results, I can see that there are two Kubernetes metrics that have never been queried (query count is 0):*\n >\n > - *`kube_persistentvolumeclaim_deletion_timestamp` - This metric provides information about when PersistentVolumeClaims were marked for deletion.*\n > - *`kube_pod_init_container_status_last_terminated_reason` - This metric shows the reason why init containers in pods were last terminated.*\n >\n > *These metrics might be valuable for monitoring and troubleshooting purposes:*\n >\n > - *The PVC deletion timestamp metric could help track deletion patterns or identify stuck PVCs.*\n > - *The init container termination reason metric could help identify common failure patterns in your initialization containers.*\n >\n > *Would you like me to explain more about what these metrics track or suggest some useful queries to start monitoring them?*\n- > **You**: Yes, please. And also, can you create relabeling rules to stop ingesting these metrics?\n- > **Assistant**: *I'll create a relabel configuration to exclude these never-queried metrics from collection. This will help reduce storage and processing overhead. Here's the relabel configuration you can add to your VictoriaMetrics scrape configuration:*\n >\n > ```yaml\n > relabel_configs:\n > - action: drop\n > source_labels: [__name__]\n > regex: kube_persistentvolumeclaim_deletion_timestamp|kube_pod_init_container_status_last_terminated_reason\n > ```\n >\n > *This configuration will:*\n > 1. *Look at each metric's name (the `__name__` label)*\n > 2. *Check if it matches either of the specified metrics*\n > 3. *Drop the metrics that match, preventing them from being collected and stored*\n >\n > *You can add this to your VictoriaMetrics configuration file or to the relevant scrape job configurations in your Prometheus configuration if you're using Prometheus as a scraper for VictoriaMetrics.*\n > *If you want to test this relabel configuration before implementing it, you can use the metric relabel debug tool to see how it would affect your metrics.*\n- > **You**: Yes, please debug this relabel config on real time series to see how it would affect my metrics\n- > **Assistant**: *Perfect! The relabel config works as expected:*\n >\n > - *The first two test cases show that both `kube_persistentvolumeclaim_deletion_timestamp` and `kube_pod_init_container_status_last_terminated_reason` metrics are correctly dropped (resulting labels are empty `{}`).*\n > - *The third test case confirms that other metrics like `kube_pod_status_phase` are preserved and not affected by our drop rule.*\n >\n > *You can confidently add this relabel configuration to your VictoriaMetrics or Prometheus configuration to stop collecting these unused metrics. This will save storage space and reduce processing overhead without affecting your current monitoring capabilities.*\n- > **You**: But i'm using vmagent resource of VictoriaMetrics operator, could you please give me part of specification for my agent with this relabel config?\n- > **Assistant**: *You're right, since you're using the VictoriaMetrics Operator and the VMAgent custom resource, the configuration needs to be in the appropriate format for that CR. Here's how you can add the relabel configuration to your VMAgent specification:*\n >\n > ```yaml\n > apiVersion: operator.victoriametrics.com/v1beta1\n > kind: VMAgent\n > metadata:\n > name: your-vmagent-name\n > spec:\n > # ... your existing configuration ...\n > \n > # Add this section to configure global metric relabeling\n > inlineRelabelConfig:\n > - action: drop\n > sourceLabels: [\"__name__\"]\n > regex: kube_persistentvolumeclaim_deletion_timestamp|kube_pod_init_container_status_last_terminated_reason\n > \n > # ... rest of your configuration ...\n > ```\n >\n > *Key points:*\n >\n > - *In the VMAgent spec, you use `inlineRelabelConfig` for inline relabeling rules*\n > - *The field is called `sourceLabels` (not `source_labels` as in Prometheus config)*\n > - *This is a global relabeling that will be applied to all metrics before they're stored*\n- > **You**: I decided not to remove those metrics, but just lower the sample rate for them using Downsampling. What downsampling parameters i have now?\n- > **Assistant**: *I can see that you currently have the following downsampling configuration:*\n >\n > ```\n > -downsampling.period=\"7d:1m,30d:5m,180d:30m\"\n > ```\n >\n > *This means:*\n >\n > - *Data newer than 7 days is kept at its original resolution*\n > - *Data between 7-30 days old is downsampled to 1-minute resolution*\n > - *Data between 30-180 days old is downsampled to 5-minute resolution*\n > - *Data older than 180 days is downsampled to 30-minute resolution*\n >\n > *Now, if you want to add specific downsampling rules for those rarely-queried metrics, you can use the downsampling filters debug tool to test how new rules would work. Let's test a configuration where we apply more aggressive downsampling to those specific metrics:*\n >\n > *...getting series and using downsampling filter debugger...*\n >\n > *Great! Now our test is working correctly. Based on the current configuration and our testing, here's a suggested downsampling configuration for your VMAgent CR that would apply more aggressive downsampling to those rarely-queried metrics:*\n >\n > ```yaml\n > apiVersion: operator.victoriametrics.com/v1beta1\n > kind: VMCluster\n > metadata:\n > name: monitoring-vm-cluster\n > spec:\n > # ... your existing configuration ...\n > \n > # Update the vmstorage section with the new downsampling parameters\n > vmstorage:\n > extraArgs:\n > downsampling.period: '{__name__=\\\"kube_persistentvolumeclaim_deletion_timestamp\\\"}:1d:10m,{__name__=\\\"kube_pod_init_container_status_last_terminated_reason\\\"}:1d:10m,7d:1m,30d:5m,180d:30m'\n > \n > # ... rest of your configuration ...\n > ``` \n >\n > *This configuration:*\n >\n > *- Applies 10-minute resolution downsampling for the rarely-queried metrics after just 1 day*\n > *- Maintains your existing downsampling policy for all other metrics*\n >\n > *The order of the downsampling rules is important - more specific rules should be listed first.*\n\nDuring this dialog, the assistant was using the following tools:\n\n- `tenants` to get the list of available tenants\n- `documentation` to get information about functionality and data formats\n- `metrics` to get the list of available metrics\n- `metrics_staistics` to get the information about metrics usage\n- `series` to get the time series for debugging\n- `metric_relabel_debug` to debug relabeling rules\n- `flags` to get the information about instance parameters\n- `downsampling_filters_debug` to debug downsampling configuration\n\nBut you can use any other tools and combine them in your own way.\n\n## Monitoring\n\nIn [SSE and HTTP modes](#modes) the MCP Server provides metrics in Prometheus format (see [endpoints](#endpoints)) and you can find [in repo simple grafana dashboard](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/blob/main/dashboard/mcp-victoriametrics-grafana.json) for these metrics. \n\n## Roadmap\n\n- [x] Support \"Prettify query\" tool (done in [`v0.0.5`](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases/tag/v0.0.5))\n- [x] Support \"Explain query\" tool (done in [`v0.0.6`](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases/tag/v0.0.6))\n- [x] Support CI pipeline for building and pushing multiarch docker images (done in [`v1.0.0`](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases/tag/v1.0.0))\n- [ ] Support tool for analysis of [Query execution statistics](https://docs.victoriametrics.com/victoriametrics/query-stats/)\n- [ ] Support vmanomaly\n- [x] Support tool for [unit-testing of alerting and recording rules](https://docs.victoriametrics.com/victoriametrics/vmalert-tool/) (done in [`v0.0.7`](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases/tag/v0.0.7))\n- [x] Support optional integration with [VictoriaMetrics Cloud](https://victoriametrics.com/products/cloud/) (via [API keys](https://docs.victoriametrics.com/victoriametrics-cloud/api/)) (done in [`v0.0.9`](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases/tag/v0.0.9))\n- [ ] Add some extra knowledge to server in addition to current documentation tool:\n - [x] [VictoriaMetrics blog](https://victoriametrics.com/blog/) posts (done in [`v1.1.0`](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases/tag/v1.1.0))\n - [ ] Github issues\n - [ ] Public slack chat history\n - [ ] CRD schemas\n - [ ] Alerting and recording rule sets\n- [ ] Implement multitenant version of MCP (that will support several deployments)\n- [ ] Add flags/configs validation tool\n- [ ] Support tools for vmagent API\n- [ ] Support [new vmalert API](https://github.com/VictoriaMetrics/VictoriaMetrics/pull/9046/files)\n- [x] Enabling/disabling tools via configuration (done in [`v0.0.8`](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/releases/tag/v0.0.8))\n- [ ] Tools for Alertmanager APIs [#6](https://github.com/VictoriaMetrics-Community/mcp-victoriametrics/issues/6)\n- [ ] Support for [metrics metadata](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/2974) in case of implementation in VictoriaMetrics\n- [ ] Support authentication\n- [ ] Add static index page with description and links to documentation\n\n## Mentions\n\n[ $\"How$ ](https://www.youtube.com/watch?v=1k7xgbRi1k0)\n\n[ $\"MCP$ ](https://victoriametrics.com/blog/q2-2025-whats-new-victoriametrics-cloud/)\n\n## Disclaimer\n\nAI services and agents along with MCP servers like this cannot guarantee the accuracy, completeness and reliability of results.\nYou should double check the results obtained with AI.\n\nThe quality of the MCP Server and its responses depends very much on the capabilities of your client and the quality of the model you are using.\n\n## Contributing\n\nContributions to the MCP VictoriaMetrics project are welcome! \n\nPlease feel free to submit issues, feature requests, or pull requests.\n","url","https://github.com/VictoriaMetrics-Community/mcp-victoriametrics","githubRepository",{"_156":157,"_158":159,"_160":161,"_16":162,"_163":164,"_31":30,"_112":165,"_167":168,"_169":170,"_172":173},"defaultBranch","main","fullName","VictoriaMetrics-Community/mcp-victoriametrics","githubTreeId",1892573,13103,"language","Go",{"_116":117,"_94":166},"Vp5XG7Qx4B","recentStargazers",0,"spdxLicense",{"_31":171},"Apache 2.0","stargazers",78,"glamaJson","inspectable",false,"npmPackage","supportedPlatforms",[180,181,182],"LINUX","WINDOWS","MACOS","scores",{"_185":186,"_187":-5,"_188":-5},"license",100,"quality","security","simpleIcon","toolCount","4ogwgtlwc2","updatedAt","2025-10-07T00:30:34.616634Z","meta",[196,199,201,208,211,213,216,219,222,225,228,230,233,236,238,240,242,244,246],{"_197":198},"title","mcp-victoriametrics | Glama",{"_200":30,"_31":29},"content",{"_202":203,"_204":205,"_206":207},"href","https://glama.ai/mcp/servers/@VictoriaMetrics-Community/mcp-victoriametrics","rel","canonical","tagName","link",{"_200":30,"_209":210},"property","og:title",{"_200":30,"_209":212},"og:description",{"_200":214,"_209":215},"https://glama.ai/logo.png","og:logo",{"_200":217,"_209":218},"https://glama.ai/mcp/servers/@VictoriaMetrics-Community/mcp-victoriametrics/og-image","og:image",{"_200":220,"_209":221},"630","og:image:height",{"_200":223,"_209":224},"1200","og:image:width",{"_200":226,"_209":227},"website","og:type",{"_200":203,"_209":229},"og:url",{"_200":231,"_209":232},"Glama – MCP Hosting Platform","og:site_name",{"_200":234,"_31":235},"summary_large_image","twitter:card",{"_200":231,"_31":237},"twitter:title",{"_200":30,"_31":239},"twitter:description",{"_200":217,"_31":241},"twitter:image",{"_200":220,"_31":243},"twitter:image:height",{"_200":223,"_31":245},"twitter:image:width",{"_247":248},"script:ld+json",{"_249":250,"_251":252},"@context","http://schema.org","@graph",[253],{"_249":254,"_255":256,"_257":258},"https://schema.org","@type","BreadcrumbList","itemListElement",[259,266,270,274],{"_255":260,"_261":262,"_31":263,"_264":265},"ListItem","item","/","Glama","position",1,{"_255":260,"_261":267,"_31":268,"_264":269},"/mcp","MCP",2,{"_255":260,"_261":271,"_31":272,"_264":273},"/mcp/servers","Servers",3,{"_255":260,"_261":275,"_31":30,"_264":276},"/mcp/servers/@VictoriaMetrics-Community/mcp-victoriametrics",4,"visitorIsMaintainer","routes/_public/mcp/servers/~namespace/~slug/_pages/_explorer/blob/~commit/~/_route",{"_280":281,"_27":287,"_314":315,"_194":329,"_326":353,"_354":355},"blob",{"_200":282,"_283":284,"_285":286},"\n \n \n \n \n \n \n \n \n \n","fileName","QuickTemplate.xml","size",927,{"_71":72,"_13":288,"_80":30,"_81":30,"_16":82,"_83":289,"_31":30,"_110":296,"_118":-5,"_119":-5,"_120":299,"_183":313,"_189":-5,"_91":30,"_190":168,"_94":191,"_192":193},[74,75,76,77,78,79],[290,292,294],{"_86":291,"_29":93,"_94":95},{"_88":89,"_31":90,"_91":92},{"_86":293,"_29":101,"_94":102},{"_88":98,"_31":99,"_91":100},{"_86":295,"_29":108,"_94":109},{"_88":105,"_31":106,"_91":107},{"_112":297,"_91":117},{"_114":298},{"_116":117},{"_122":300,"_149":308,"_154":309,"_174":-5,"_175":176,"_177":-5,"_178":312},{"_124":301,"_146":307,"_131":148},{"_126":302,"_133":303,"_137":304,"_140":305,"_143":306},{"_128":129,"_29":130,"_131":132},{"_128":135,"_29":136,"_131":132},{"_29":139,"_131":132},{"_29":142,"_131":132},{"_29":145,"_131":132},[140,137],{"_32":151,"_152":153},{"_156":157,"_158":159,"_160":161,"_16":162,"_163":164,"_31":30,"_112":310,"_167":168,"_169":311,"_172":173},{"_116":117,"_94":166},{"_31":171},[180,181,182],{"_185":186,"_187":-5,"_188":-5},"member",{"_316":317,"_320":321,"_322":323,"_160":324,"_325":176,"_326":284,"_327":328,"_285":286,"_131":280},"commit",{"_318":319},"sha","88c50faf046539103cd2deac0a09ae47dff850b4","contentStoredObjectPointerId",14003057,"fullPath","vendor/github.com/valyala/quicktemplate/QuickTemplate.xml",1516081,"indexable","path","safeToDisplayAsText",true,[330,332,334,336,338,339,340,342,343,344,345,346,347,348,349,350,351,352],{"_197":331},"QuickTemplate.xml | mcp-victoriametrics | Glama",{"_200":333,"_31":29},"Enterprise-grade security, privacy, and features like agents, MCP, prompt templates, and more.",{"_202":335,"_204":205,"_206":207},"https://glama.ai/mcp/servers/@VictoriaMetrics-Community/mcp-victoriametrics/blob/88c50faf046539103cd2deac0a09ae47dff850b4/vendor/github.com/valyala/quicktemplate",{"_200":337,"_209":210},"QuickTemplate.xml | mcp-victoriametrics",{"_200":333,"_209":212},{"_200":214,"_209":215},{"_200":341,"_209":218},"https://glama.ai/generated-images/og?title=QuickTemplate.xml+%7C+mcp-victoriametrics",{"_200":220,"_209":221},{"_200":223,"_209":224},{"_200":226,"_209":227},{"_200":335,"_209":229},{"_200":231,"_209":232},{"_200":234,"_31":235},{"_200":231,"_31":237},{"_200":333,"_31":239},{"_200":341,"_31":241},{"_200":220,"_31":243},{"_200":223,"_31":245},"vendor/github.com/valyala/quicktemplate","treeMembers",[356,360,364,368,372,376,380,382,386,390,394,398,402,406],{"_316":357,"_322":358,"_326":359,"_131":280},{"_318":319},"vendor/github.com/valyala/quicktemplate/bytebuffer.go","bytebuffer.go",{"_316":361,"_322":362,"_326":363,"_131":280},{"_318":319},"vendor/github.com/valyala/quicktemplate/doc.go","doc.go",{"_316":365,"_322":366,"_326":367,"_131":280},{"_318":319},"vendor/github.com/valyala/quicktemplate/.gitignore",".gitignore",{"_316":369,"_322":370,"_326":371,"_131":280},{"_318":319},"vendor/github.com/valyala/quicktemplate/htmlescapewriter.go","htmlescapewriter.go",{"_316":373,"_322":374,"_326":375,"_131":280},{"_318":319},"vendor/github.com/valyala/quicktemplate/jsonstring.go","jsonstring.go",{"_316":377,"_322":378,"_326":379,"_131":280},{"_318":319},"vendor/github.com/valyala/quicktemplate/LICENSE","LICENSE",{"_316":381,"_322":323,"_326":284,"_131":280},{"_318":319},{"_316":383,"_322":384,"_326":385,"_131":280},{"_318":319},"vendor/github.com/valyala/quicktemplate/README.md","README.md",{"_316":387,"_322":388,"_326":389,"_131":280},{"_318":319},"vendor/github.com/valyala/quicktemplate/.travis.yml",".travis.yml",{"_316":391,"_322":392,"_326":393,"_131":280},{"_318":319},"vendor/github.com/valyala/quicktemplate/urlencode.go","urlencode.go",{"_316":395,"_322":396,"_326":397,"_131":280},{"_318":319},"vendor/github.com/valyala/quicktemplate/util_appengine.go","util_appengine.go",{"_316":399,"_322":400,"_326":401,"_131":280},{"_318":319},"vendor/github.com/valyala/quicktemplate/util.go","util.go",{"_316":403,"_322":404,"_326":405,"_131":280},{"_318":319},"vendor/github.com/valyala/quicktemplate/util_noappengine.go","util_noappengine.go",{"_316":407,"_322":408,"_326":409,"_131":280},{"_318":319},"vendor/github.com/valyala/quicktemplate/writer.go","writer.go","actionData","errors"]

MCP directory API