io.github.ryanmat/logicmonitor
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| AWX_URL | No | Ansible Automation Platform controller URL (e.g., `https://aap.example.com`) | |
| AWX_TOKEN | No | AAP personal access token | |
| HF_DEVICE | No | Torch device for inference (cpu, cuda, mps, auto) | auto |
| LM_PORTAL | Yes | LogicMonitor portal hostname (e.g., `company.logicmonitor.com`) | |
| LM_AGE_KEY | No | Path to the age identity that decrypts the vault (multi-portal) | |
| LM_TIMEOUT | No | Request timeout in seconds (range: 5-300) | 30 |
| TF_TIMEOUT | No | Terraform command timeout in seconds | 300 |
| AWX_TIMEOUT | No | Request timeout in seconds for AAP API calls | 30 |
| WATSONX_URL | No | IBM watsonx.ai endpoint URL | https://us-south.ml.cloud.ibm.com |
| HF_CACHE_DIR | No | HuggingFace model cache directory | |
| HF_LLM_MODEL | No | HuggingFace LLM model name or path | ibm-granite/granite-3.3-2b-instruct |
| HF_TTM_MODEL | No | HuggingFace TTM model name or path | ibm-granite/granite-timeseries-ttm-r2 |
| LM_ACCESS_ID | No | LMv1 API access ID (for ingestion APIs) | |
| LM_HTTP_HOST | No | HTTP server bind address | 0.0.0.0 |
| LM_HTTP_PORT | No | HTTP server port | 8080 |
| LM_LOG_LEVEL | No | Logging level: `debug`, `info`, `warning`, or `error` | warning |
| LM_TRANSPORT | No | Transport mode: `stdio` (local) or `http` (remote) | stdio |
| LM_ACCESS_KEY | No | LMv1 API access key (for ingestion APIs) | |
| LM_VAULT_FILE | No | Path to the age-encrypted portal vault (multi-portal) | |
| AWX_VERIFY_SSL | No | Verify SSL certificates for AAP connections | true |
| LM_API_VERSION | No | API version | 3 |
| LM_MAX_RETRIES | No | Max retries for rate-limited/server error requests (range: 0-10) | 3 |
| AWX_MAX_RETRIES | No | Max retries for failed AAP API requests | 3 |
| LM_BEARER_TOKEN | No | API Bearer token (min 10 characters) | |
| LM_CORS_ORIGINS | No | Comma-separated CORS origins (default: none) | |
| LM_MULTI_PORTAL | No | Serve many customer portals from one server, selected at runtime via `use_portal`. Stdio-only. | false |
| LM_PORTALS_FILE | No | Plaintext JSON portal map (multi-portal, testing only; the encrypted vault wins when both are set) | |
| WATSONX_API_KEY | No | IBM Cloud API key for watsonx.ai (enables Granite TTM + NL summaries) | |
| WATSONX_TIMEOUT | No | Request timeout in seconds for watsonx.ai API calls | 60 |
| LM_ENABLED_TOOLS | No | Comma-separated tool names or glob patterns to enable (e.g., `get_*,triage`). Mutually exclusive with `LM_DISABLED_TOOLS`. | |
| TF_WORKSPACE_DIR | No | Root directory for Terraform workspaces (enables Terraform tools) | |
| LM_DISABLED_TOOLS | No | Comma-separated tool names or glob patterns to disable (e.g., `delete_*`). Mutually exclusive with `LM_ENABLED_TOOLS`. | |
| LM_MCP_CATEGORIES | No | Comma-separated category names to include: `read`, `write`, `delete`, `export`, `import`, `session`, `workflow`. Composes by intersection with `LM_ENABLED_TOOLS`/`LM_DISABLED_TOOLS` -- only narrows, never expands. Useful for clients with tool-count limits (e.g., Cursor's 40-tool cap). | |
| LM_HTTP_AUTH_TOKEN | No | Require this bearer token on `/mcp` and `/api/v1/*` (min 16 chars). Health endpoints and `/` stay open for probes. HTTP transport only. | |
| LM_SESSION_ENABLED | No | Enable session context tracking | true |
| WATSONX_PROJECT_ID | No | IBM watsonx.ai project ID | |
| LM_FIELD_VALIDATION | No | Field validation: `off`, `warn`, or `error` | warn |
| TF_TERRAFORM_BINARY | No | Path to the terraform binary | terraform |
| LM_SESSION_HISTORY_SIZE | No | Number of tool calls to keep in history | 50 |
| LM_SESSION_PERSIST_PATH | No | File path for persistent session variables (survives restarts) | |
| TF_AUTO_APPROVE_ENABLED | No | Enable terraform apply/destroy operations | false |
| LM_ENABLE_WRITE_OPERATIONS | No | Enable write operations (create, update, delete) | false |
| LM_HEALTH_CHECK_CONNECTIVITY | No | Include LM API ping in health checks | false |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server exposes up to 273 LogicMonitor tools, more than most clients load at once. To find the right tool for a task, call search_tools with relevant keywords (or a category) first instead of enumerating the full list. The composite workflow tools (triage, diagnose, health_check, portal_overview, capacity_plan and detect_site_outage) answer common multi-step investigations in a single call. Tools that change the portal (create, update, delete, acknowledge, run) are listed only when the operator sets LM_ENABLE_WRITE_OPERATIONS=true.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| get_devicesB | List devices (resources) from LogicMonitor, optionally filtered by name, hostname, status or device group |
| get_deviceB | Get detailed information about a specific device (resource) |
| get_device_groupsC | List device/resource groups from LogicMonitor |
| get_device_groupA | Get detailed information about a specific device/resource group, including appliesTo expression and parent ID |
| get_device_deltaA | List devices/resources through the device delta, to follow what changed. Without delta_id it lists the devices whose display name contains name_filter, or every device, and returns a delta_id; each such read starts a new delta, so keep the first page's. With delta_id it lists the devices changed since that delta started (change says how), every change since then on each read. name_filter is refused there, since the portal answered a filtered read with an empty page that left out a changed device. |
| get_unmonitored_devicesB | List the portal's unmonitored devices/resources, found by a NetScan and not monitored, rows as the API returns them. |
| get_alertsA | Get open or cleared alerts from LogicMonitor, optionally filtered by severity (critical, error, warning), device, group, datasource or time For Kubernetes clusters, use group_id (from get_device_groups) instead of device; the device name filter does not work reliably for K8s resources. Common mistakes: startEpoch/endEpoch use SECONDS not milliseconds. String filter values need double quotes (e.g., monitorObjectName:"hostname"). |
| get_alert_detailsB | Get detailed information about a specific alert |
| get_sdtsC | List scheduled downtimes from LogicMonitor |
| get_active_sdtsB | Get currently active SDTs |
| get_upcoming_sdtsA | Get SDTs scheduled to start within a time window |
| get_collectorsC | List collectors from LogicMonitor |
| get_collectorA | Get detailed information about a specific collector |
| get_collector_groupsB | List collector groups |
| get_collector_groupA | Get details about a specific collector group |
| get_collector_eventsA | List one collector's events, newest first: agent and sbproxy restarts, credential rotation, and reports that the collector is unreachable, each with its time and build. Use it for a collector's recent history; get_collector_health gives its current state. |
| get_collector_log_filesA | List the log files a collector has uploaded to the portal, as zip files with their size and upload time. A collector uploads them through Send Logs to LogicMonitor in the portal or the !uploadlog debug command, so an empty list means nothing was uploaded. Read one with get_collector_log_file. |
| get_collector_log_fileA | Read one uploaded collector log file, named as get_collector_log_files lists it, and return the end of its text: at most max_bytes bytes, starting on a whole line. truncated says whether earlier lines were left out. |
| get_collector_upgrade_historyA | List the portal's collector version upgrades a page at a time (limit, offset), each row as the API returns it. |
| get_device_datasourcesA | Get datasources applied to a device (resource) Common mistakes: Returns device-datasource associations not definitions. The ID returned here is device_datasource_id for use with get_device_data. |
| get_device_instancesC | Get instances of a datasource on a device (resource) |
| get_device_dataA | Get the raw metric values (samples) of a device/resource datasource instance, such as CPU, memory or disk usage Common mistakes: Returns the last hour unless start_time or end_time is given. Requires device_datasource_id (from get_device_datasources) not the datasource definition ID. |
| get_graph_dataC | Get graph image data for visualization |
| get_dashboardsB | List dashboards from LogicMonitor |
| get_dashboardC | Get detailed information about a specific dashboard |
| get_dashboard_widgetsB | Get widgets configured on a dashboard |
| get_widgetB | Get details about a specific widget |
| get_dashboard_groupsC | List dashboard groups |
| get_dashboard_groupB | Get details about a specific dashboard group |
| get_websitesC | List websites from LogicMonitor |
| get_websiteB | Get detailed information about a specific website |
| get_website_groupsC | List website groups |
| get_website_checkpointsA | List the portal's website checkpoint locations, the places website checks run from; each id is the smg_id a website's checkpoints carry in get_website. |
| get_website_dataC | Get synthetic check data for a website |
| get_reportsC | List reports from LogicMonitor |
| get_reportB | Get detailed information about a specific report |
| get_report_groupsC | List report groups |
| get_scheduled_reportsB | Get reports with schedules configured |
| get_report_executionA | Poll the status of a report generation task started by run_report (returns status and result URL when finished) |
| get_escalation_chainsC | List escalation chains |
| get_escalation_chainC | Get details about a specific escalation chain |
| get_recipient_groupsC | List recipient groups |
| get_recipient_groupB | Get details about a specific recipient group |
| get_integrationsA | List LogicMonitor integrations (Custom HTTP Delivery, Slack, PagerDuty, etc.). Returns a short summary per integration. |
| get_integrationA | Get a specific integration's full definition. Field set depends on integration type; password/OAuth secret fields are masked. |
| get_alert_settingsA | Read datapoint alert thresholds at instance, device or group level: each datapoint's threshold in force and the level it comes from (instance, instance_group, device_datasource, device_group with source_group, group or global), beside the level's own and the global threshold. Use it before tuning a noisy threshold with update_alert_settings. |
| get_alert_rulesB | List alert rules: each rule's priority and the escalation chain it routes matching alerts to |
| get_alert_ruleC | Get details about a specific alert rule |
| get_usersA | List users from LogicMonitor |
| get_userA | Get details about a specific user |
| get_rolesB | List roles |
| get_roleA | Get details about a specific role |
| get_access_groupsC | List access groups |
| get_access_groupA | Get details about a specific access group |
| get_api_tokensC | List API tokens |
| get_api_tokenB | Get details about a specific API token |
| get_tenantsA | List the portal's tenant names, as the tenants API answers them: names only, with no paging or detail. |
| get_device_propertiesA | Get all properties of a device (resource). Each has its value, its type (custom, inherit, auto or system) and |
| get_device_propertyA | Get a specific property of a device (resource): its value, its type (custom, inherit, auto or system) and |
| get_device_group_propertiesA | Get a device group's (resource group's) properties: its own, the ones it inherits and the system ones, each with the value in force, its type (custom, inherit or system) and the ancestor groups that hold the same name, nearest first; an inherited value is the nearest one's. property_name reads one property. |
| get_datasourcesC | List datasources from LogicMonitor |
| get_datasource_devicesA | List the devices/resources a datasource applies to, each with up to 10 of its instances; more_instances says it has others, which get_device_instances lists. |
| get_datasourceC | Get details about a specific datasource |
| get_configsourcesC | List ConfigSources |
| get_configsourceC | Get details about a specific ConfigSource |
| get_configsource_update_reasonsB | Get update history and audit trail for a ConfigSource |
| get_device_configB | List config versions collected for a device instance |
| get_device_config_versionA | Get a specific config version with full content and diffs |
| get_eventsourcesB | List EventSources |
| get_eventsourceC | Get details about a specific EventSource |
| get_device_eventsourcesB | Get EventSources applied to a device (resource). Returns device-level EventSource associations. |
| get_propertysourcesC | List PropertySources |
| get_propertysourceC | Get details about a specific PropertySource |
| get_topologysourcesC | List TopologySources |
| get_topologysourceC | Get details about a specific TopologySource |
| get_logsourcesC | List LogSources |
| get_logsourceC | Get details about a specific LogSource |
| get_device_logsourcesC | Get LogSources applied to a device (resource) |
| get_netscansC | List network scans |
| get_netscanB | Get details about a specific network scan |
| get_oidsC | List OID definitions |
| get_oidC | Get details about a specific OID |
| get_servicesB | List Service Insight business services (deviceType 6 devices, including APM trace services) |
| get_serviceA | Get details about a specific Service Insight service |
| get_service_groupsB | List Service Insight service groups (BizService device groups) |
| get_service_insights_rulesA | List the portal's Service Insights rules, rows as the API returns them. offset must be a multiple of limit; the portal answers any other offset with a 400. get_services lists the services themselves. |
| get_ops_notesC | List ops notes |
| get_ops_noteC | Get details about a specific ops note |
| get_audit_logsA | Read the portal audit log: call this for who did what and when. kind='all' returns one page of entries, newest first; kind='change' returns the configuration changes (create, update, delete or import, read from each entry's description) found in a pool of the newest entries, with scanned and has_more; kind='login' returns portal sign-ins, sign-outs and failed logins, one query per login wording (up to six calls), with status and method per row, success_only or failed_only, and total as a lower bound while has_more is true; kind='api_token' returns the rows naming an API token, given by access_id or token_id (read to its access ID), keeping only that exact access ID, with event (change, rejected, throttled), path and method per row. username is an exact name or a comma-separated list; action and resource_type (kinds all and change) are whole words the description holds (the portal applies one, the second is matched on the fetched page); start_time and end_time bound the entry time in epoch seconds. System and automation actors (System:*, LogicMonitor Internal, cloud auto-delete) are dropped after the fetch for kinds all and change unless exclude_system is false or a username is given; dropped counts them and paging follows the rows the portal returned. An argument another kind reads is refused. offset plus the rows read must stay within 10,000, and the filter within about 900 characters. |
| get_topology_mapC | Get network topology map data |
| get_device_neighborsB | Get neighboring devices/resources based on topology |
| get_device_interfacesC | Get network interfaces for a device (resource) |
| get_device_connectionsC | Get device/resource relationships and connections |
| get_batchjobsC | List batch jobs |
| get_batchjobA | Get details about a specific batch job |
| get_device_batchjobsA | List BatchJob datasources applied to a device (resource); per-run output lives in instance data via get_device_data |
| get_scheduled_downtime_jobsC | Get batch jobs related to SDT automation |
| get_usage_contractA | Read the portal's contracted entitlements, for license and contract questions: every feature the contract names, entitled or not, each with its role (billable_line, metered_no_entitlement, metered_unpaid, repeats_parent, rollup_or_carveout), entitled quantity, unit and roll-up basis, plus the detected billing_mode. It does not count the portal's current resources. |
| get_usage_metricsA | Read how many devices, Kubernetes, AWS, Azure and GCP resources, websites and services the portal holds now, as point-in-time counts. The portal is billed on each feature's roll-up basis in the contract, so these are not billed figures. |
| get_cost_recommendationsB | Get cost optimization recommendations. Category filter takes the category description string from get_cost_recommendation_categories |
| get_cost_recommendation_categoriesA | Get cost recommendation categories with counts and savings |
| get_cost_recommendationA | Get a specific cost recommendation by ID (v224 API) |
| get_otlp_metric_namesA | List the names of metrics ingested through native OpenTelemetry (OTLP), optionally those containing a substring, to find what to query with query_otlp_metrics or analyze_metric(source="otlp"). A portal without OTLP Metrics answers with available: false |
| get_otlp_metric_labelsA | List the label names on native OTLP metrics, optionally narrowed to one metric or to names containing a substring |
| get_otlp_label_valuesA | List the values observed for one native OTLP metric label, such as every service_name, optionally narrowed to one metric |
| query_otlp_metricsA | Run a PromQL query against native OTLP metrics. A range query (start and end in epoch seconds, optional step) returns each series' samples over the window; instant=true returns each series' value now and takes no window. Returns at most 20 series and 500 points a series, downsampled; for statistics on a series use analyze_metric(source="otlp") |
| get_log_pipelinesA | List every LM Logs pipeline with its log query and the partitions it names. It returns all pipelines at once and takes no arguments. It does not search log content. |
| get_tracked_log_queriesA | List the portal's tracked log queries, each row as the API returns it. Use it to see which log queries the portal tracks; it does not run them. |
| get_log_query_groupsB | List the saved LM Logs query groups (folders) with their type, owner, permission and the queries each holds. |
| get_log_partitionsA | List the LM Logs partitions with whether each is active or paused, its criteria, and its active contract (state, SKU, retention, usage limit, period). Use it to check partition state and contract limits. |
| get_log_usageA | Read the portal's LM Logs usage figure from /log/usage: one number, as returned. The API states no unit or period for it. |
| analyze_alertsA | Analyze one alert window: call this for how many alerts fired and when, which alerts belong together, how noisy alerting is, which datasources burst, or which portal changes preceded a spike over the last hours. One read of the window's active and cleared alerts, under the severity, device, group and datasource filters, feeds every section; 'statistics' gives active and cleared totals, counts by severity, the ten busiest devices and datasources, and the volume of each hour, cleared alerts included; 'clusters' groups two or more alerts by device, by datasource and by signature (datasource and datapoint) and lists the five largest of each type, each with its alert count, its ten most severe alert ids, up to ten of its devices with device_count, and first and last start times, while cluster_count and by_type count every cluster; a device cluster names the device_id its alerts share, null for a website or two devices under one name; 'noise' gives a 0 to 100 score from the entropy of alerts over datasource and datapoint (weight 40), flaps, an alert on the same device and datapoint that re-fires within 30 minutes of clearing (30), and repeats, a combination with three or more alerts (30), with the first ten flap pairs and tuning recommendations; 'bursts' gives the windows of window_seconds on one datasource that hold at least min_alerts alerts on min_devices devices, the ten largest listed and all counted in bursts_detected; 'changes' reads the audit log once and pairs each change (a create, update, delete or import, read from the entry's description) with the first 5-minute alert spike, a bucket above the mean by one standard deviation, within correlation_window_minutes after it, with a confidence from 1.0 to 0.5; it lists the ten pairs of highest confidence and counts them all in total_correlated. Narrow the filters to see what a bounded list leaves out. truncated says when the 5000-alert cap cut either read short. Use get_alerts to list or page through the alerts themselves. |
| correlate_metricsA | Compute Pearson correlation between multiple metric series. Joins the series by timestamp bucket, as wide as the coarsest series' median sample interval with finer series averaged into it, and correlates over the buckets every series holds, reported as shared_samples. Builds an NxN correlation matrix and highlights strong correlations (|r| > 0.7). Maximum 10 sources. |
| calculate_availabilityA | Calculate availability percentage from alert history. Computes SLA-style uptime metrics, MTTR, and per-device breakdown from the window's cleared and active alerts, read as two lists; truncated says when the 5000-alert cap cut either read short. With target_slo it adds error_budget for the window: budget, consumed and remaining minutes, percent consumed, burn rate, projected exhaustion in days and a status (healthy, warning past 50 percent, critical past 90, exhausted). Common mistakes: hours_back defaults to 720 (30 days). Narrow scope with device_id/group_id for performance. |
| analyze_blast_radiusA | Count the devices a device failure could reach through topology neighbors: neighbor_count at depth 1, dependent_count beyond it, each device's alert status, and critical path devices. A device whose neighbors neither topology path answers is listed under unreadable with a warning, never counted as zero; on a portal without those routes every read is unreadable. |
| analyze_metricA | Analyze one device instance's metric series, such as CPU, memory or disk usage: call this for statistics on a datapoint's recent history, such as its current level against the window, whether it is rising, its daily cycle, when it shifted, or which samples stand out. One paged read of the window feeds every section; 'summary' gives latest, mean, standard deviation, min, max and the latest sample's signed z-score; 'trend' gives least-squares and Theil-Sen slopes per hour, r-squared, coefficient of variation, the detrended 24-hour autocorrelation and a label (volatile, cyclic, increasing, decreasing or stable); 'seasonality' gives detrended autocorrelation at 1, 4, 12, 24 and 168 hours, the dominant period and the peak hours in UTC; 'change_points' gives level shifts, where the mean of the next 30 samples leaves the mean of the previous 30 by five times the noise in that comparison (measured with any steady trend and autocorrelation the noise has), with the daily cycle removed first when the window holds three days or more and that leaves less noise; each shift has its timestamp and the levels before and after it, less the daily cycle when that was removed. A change in the last 30 samples is not reported yet, small shifts in strongly autocorrelated noise go unfound, on a window under a day a change inflates the noise and smaller shifts go unfound, and changes about a day apart in a few days leave the cycle in and go unfound; 'anomalies' gives z-score, IQR and MAD hits, each method's 20 most extreme; 'forecast' gives the least-squares and Holt-Winters (daily season, two days of samples needed) slopes per hour and, with threshold, when each crosses it, within 30 days; 'baseline' compares the window with one as long ending baseline_hours_ago hours ago (one more read): both windows' statistics, the mean's percent deviation and a status (normal within 20 percent, elevated or reduced within 50, anomalous past it). The baseline runs only with baseline_hours_ago. Use get_device_data for the raw samples. With source="otlp" it reads one PromQL range query (query, hours, step) in place of a device instance and runs the same sections on each series it returns, the first 20. |
| get_session_contextB | Get current session context (variables and history) |
| set_session_variableA | Set a user-defined session variable for use across tool calls |
| get_session_variableB | Get a user-defined session variable |
| delete_session_variableA | Delete a user-defined session variable |
| clear_session_contextB | Clear all session context (variables and history) |
| get_session_historyC | List recent tool call history |
| get_apm_servicesA | List APM trace services, the deviceType 6 devices, one page at a time. Entry point for discovering traced services; read a service's datasources, operations (the instances of its operations datasource), metrics and properties with get_device_datasources, get_device_instances, get_device_data and get_device_properties, passing the service id as device_id. |
| get_apm_service_alertsB | Get the alerts on one APM trace service, one page at a time |
| get_diagnosticsourcesB | List DiagnosticSources from LogicMonitor |
| get_diagnosticsourceB | Get details about a specific DiagnosticSource including datapoints |
| get_remediationsourcesC | List RemediationSources from LogicMonitor |
| get_remediationsourceA | Get details about a specific RemediationSource including the Groovy script |
| get_diagnostic_remediation_assignmentsA | List the diagnostic and remediation sources assigned to a specific resource or alert (Automated Diagnostics & Remediation). Unlike get_diagnosticsources/get_remediationsources, this resolves which modules actually apply to the target. |
| get_diagnostic_remediation_resultsA | Get structured execution results for diagnostic and remediation source runs: status, trigger type, executor, script output, and timing. Provide exactly one of alert_id or host_id. Time window params are epoch milliseconds; result timestamps are epoch seconds. |
| get_action_chainsB | List action chains: ordered DiagnosticSource/RemediationSource stages that action rules trigger on alerts (Automated Diagnostics & Remediation) |
| get_action_chainA | Get details about a specific action chain including its stages |
| get_action_rulesB | List action rules: alert conditions (severity, device groups, datasource matchers) that trigger action chains |
| get_action_ruleA | Get details about a specific action rule |
| triageA | Triage the alerts of the last hours_back hours in one call. Reads the alert window once through analyze_alerts for its statistics, clusters (by device, datasource and signature), noise and changes, then counts the blast radius of up to three device clusters, largest first. Each count row gives the devices reached as neighbors and as dependents, how many of them have a critical alert open, and how many devices' neighbors could not be read; a website cluster gets none. The clusters section lists the five largest clusters of each type with ten alert ids each, and changes the ten change and spike pairs of highest confidence; cluster_count, by_type and total_correlated count them all, and a narrower severity, device or window shows the rest. detail_level=full adds the devices each blast radius reached. Use analyze_alerts for bursts or a single section, and analyze_blast_radius for one device. |
| health_checkA | Composite health check: resolves a device, picks up to five of its datasources whose names mention CPU, memory or disk (drive, volume, filesystem), reads the first instance of each through analyze_metric (summary and anomalies over 24 hours), checks open alerts, and calculates 30-day availability. A device with none of those datasources gets an empty metrics list and a warning. |
| capacity_planA | Composite capacity planning: reads each instance once through analyze_metric to classify its trend, detect seasonality and change points, and forecast when it crosses threshold (default 90). Summary rows carry each forecast method's breach_eta (epoch seconds) and hours_until_breach per datapoint. |
| portal_overviewA | Shift-handoff overview of the portal: reads the alert window of the last hours_back hours once for statistics, clusters and noise, and lists the collectors and the active SDTs. alert_clusters lists the five largest clusters of each type with ten alert ids each, and its cluster_count and by_type count them all. detail_level=full adds up to 20 open critical and 20 open error alerts from across the portal, not only the window, and up to 20 dead devices; summary does not read them. |
| diagnoseA | Composite diagnosis: given an alert, or a device's latest open critical alert, gathers the alert details, device context, alerts clustered and audit changes in the last four hours, blast radius, and the device's assigned DiagnosticSources with the latest results of the alert's diagnostic runs. correlated_alerts lists the five largest clusters of each type with ten alert ids each, and changes the ten change and spike pairs of highest confidence; cluster_count, by_type and total_correlated count them all. It reads those results and never runs a DiagnosticSource. It computes no health score or root cause; health_check reports each metric's z-score. |
| search_toolsA | Search the tools this server offers under its current settings by keyword, optionally within one category. Use it to find the tool for a task. It never names a tool the settings hide, so with writes off no write tool appears. Words such as 'the' and 'on' are ignored, plurals match singulars, and a query word that is a tool's full name ranks that tool first. |
| get_referenceA | Get LogicMonitor reference content (schemas, enums, filter syntax, guides). Mirrors content from MCP Resources for clients without full Resource support (Copilot cloud agent, OpenAI Codex, Cline). Categories: schema, enums, filters, syntax, guide. Pass list=true (or omit both category and name) to discover all available (category, name) pairs. |
| get_terraform_importsA | List a resource type's portal objects as Terraform import blocks for the logicmonitor/logicmonitor provider, one |
| get_interface_metricsA | Pull interface-level metrics (in/out bytes, errors, discards, utilization, status) for a device's interface over a time window. Answers 'how is this port performing?' Resolves the Interface-family DataSource and the instance matching the interface name before fetching datapoints. |
| get_device_flowsA | List a NetFlow exporter device's flow records over the last hours_back hours, rows as the API returns them; the API reference gives each flow its source and destination IP, port, DNS name and ASN, its protocol and its bytes. Answers FEATURE_DISABLED when the portal has NetFlow off for the device. |
| get_device_top_talkersA | A NetFlow exporter device's top talkers over the last hours_back hours, for 'what is consuming my WAN?': by=endpoint pages the endpoints it saw and by=port the ports, rows as the API returns them (the API reference gives each a usage and a share); by=graph returns its top talkers graph whole. Answers FEATURE_DISABLED when the portal has NetFlow off for the device. |
| get_link_flapsA | Identify interfaces with repeated up/down transitions in a time window. Answers 'which ports are unstable?' Common causes: bad cable, duplex mismatch, bad SFP, PoE power cycling, WAN instability. |
| get_collector_healthA | Enriched collector status: whether each collector is down, downstream device count, uptime, and optional alert history. Leading indicator for site-level events. Prefer this over |
| get_power_eventsA | Filter alerts for UPS/PDU power-event signatures across APC, Liebert, and Eaton DataSources ('on battery', 'runtime remaining', 'input voltage lost') over a time window. Returns events matched by DataSource or alert name substring with counts per pattern. |
| detect_site_outageA | Composite workflow for site outage detection. Chains CollectorDown detection, mass-interface-down burst analysis, UPS on-battery events, and downstream device silence into a single site-outage verdict with confidence score, scope, and affected device list. Designed to catch the class of site-outage that generic AIOps correlation misses. Pass a device group ID representing the site. |
| audit_network_monitoring_coverageA | Portal audit that counts UPS/PDU devices onboarded, interface DataSources applied, SNMP credentials configured, and NetFlow exporters set up. Returns a prioritized gap list, each gap with the onboarding step that would let the portal detect what it cannot detect today. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| incident_triage | Analyze active alerts, identify patterns, and suggest root cause |
| capacity_review | Review resource utilization and identify capacity concerns |
| remediation | RemediationSource execution workflow with pre-execution safety checklist |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| Alert Schema | LogicMonitor alert object fields, types, and descriptions |
| Device Schema | LogicMonitor device object fields, types, and descriptions |
| SDT Schema | LogicMonitor SDT (Scheduled Downtime) object fields |
| Dashboard Schema | LogicMonitor dashboard object fields |
| Collector Schema | LogicMonitor collector object fields |
| Escalation Chain Schema | LogicMonitor escalation chain object fields |
| Report Schema | LogicMonitor report object fields |
| Website Schema | LogicMonitor website check object fields |
| DataSource Schema | LogicMonitor DataSource definition fields |
| User Schema | LogicMonitor user object fields |
| Audit Log Schema | LogicMonitor audit log entry fields |
| Alert Severity Values | Valid severity levels: critical(4), error(3), warning(2), info(1) |
| Device Status Values | Valid device status words: normal, dead, dead-collector, etc. |
| SDT Type Values | Valid SDT types: DeviceSDT, DeviceGroupSDT, etc. |
| Alert Cleared Status | Alert cleared status values: true, false |
| Alert Acknowledged Status | Alert acknowledgment status values: true, false |
| Collector Build Types | Collector build version types: EA, GD, MGD |
| Alert Filter Fields | Filter fields and operators for alert queries |
| Device Filter Fields | Filter fields and operators for device queries |
| SDT Filter Fields | Filter fields and operators for SDT queries |
| Filter Operators | LogicMonitor API filter operators: :, ~, >, <, !:, !~, >:, <: |
| Tool Categories | All LogicMonitor MCP tools organized by domain category |
| Common Query Examples | Common filter patterns and query examples for LogicMonitor API |
| MCP Server Orchestration Guide | Patterns for combining LogicMonitor with other MCP servers |
| Best Practices Guide | Scenario-based best practices for alert noise, device health, availability, and remediation |
TDQS
Scored across 150 tools
Most list/detail pairs are clearly named and many composite tools have detailed descriptions explaining their boundaries. However, with 150 tools there is substantial overlap between alert-related tools (get_alerts, analyze_alerts, triage, diagnose, portal_overview), collector tools (get_collectors, get_collector_health), and service tools (get_services, get_apm_services), making tool selection difficult despite the descriptions.
The dominant pattern is get_<resource> for retrieval, with fairly consistent singular/plural distinctions. But the set also mixes analyze_*, calculate_*, query_*, search_tools, set/delete/clear session verbs, and standalone noun-like names such as triage, diagnose, health_check, capacity_plan, and portal_overview, so conventions are only partially consistent.
150 tools is an extreme surface for a single MCP server, far beyond typical limits and likely to overwhelm an agent's context and decision-making. Many tools are narrow variants or composite shortcuts that could be consolidated.
The read/investigation surface is very broad, covering devices, alerts, collectors, datasources, metrics, logs, topology, NetFlow, cost, and governance. However, obvious dead ends exist: get_report_execution polls a task started by run_report, but run_report is not listed, and get_alert_settings references update_alert_settings, which is also absent. Core write/update operations are largely missing outside session variables.