incident-commander-mcp
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@incident-commander-mcp Investigate the error spike in payment-service around 14:03 UTC"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
incident-commander-mcp
AI-powered incident management MCP server built for the NitroStack Hackathon.
⚠️ R12 Disclosure — Simulated Integrations
All data returned by this MCP server is entirely mocked and stored in-memory.
There are NO real connections to:
Kubernetes / container orchestration platforms
Grafana / Prometheus / any metrics backend
GitHub / GitLab / any version control system
PagerDuty / OpsGenie / any alerting system
Stripe or any payment processor
This is a demo for hackathon evaluation purposes. The realistic-looking data is seeded with a purposeful incident scenario (deploy at 14:02 → error spike at 14:03) to demonstrate agent reasoning capabilities over MCP tools, resources, and prompts.
Related MCP server: mttrly
Overview
incident-commander-mcp is a single MCP server built on the NitroStack TypeScript framework
(@nitrostack/core). It exposes tools, resources, and prompts that enable an AI agent to:
Investigate production incidents by querying deployment history, error logs, and metrics
Reason about root causes using cross-correlated data
Act (with human approval) by requesting rollbacks or creating incident tickets
Document incidents via structured postmortem prompts
Capabilities
🔧 Tools (8)
Tool | Description | Annotation |
| Recent deployments, optionally filtered by service |
|
| Error logs for a service within a time window |
|
| Latency, error rate, CPU/memory time-series |
|
| Kubernetes pod readiness and restart counts |
|
| Cross-correlates deploy + logs + metrics → hypothesis |
|
| Submits a rollback request (requires human approval) |
|
| Creates a new in-memory incident record | — |
| Appends a timeline entry to an incident | — |
📦 Resources (4)
URI | Description |
| Active incident snapshot (JSON) |
| Timeline for a specific incident |
| Last 20 deployments across all services |
| Health snapshot for one service |
💬 Prompts (3)
Prompt | Arguments | Description |
|
| Guides structured RCA with tool call sequence |
|
| Generates a blameless postmortem |
|
| Risk/benefit analysis before rollback |
Incident Scenario (Demo)
The seeded data tells a realistic story:
14:00 UTC — payment-service healthy: error_rate 0.2%, latency 45ms
14:02 UTC — payment-service v2.4.1 deployed (commit a3f8c21d)
14:03 UTC — Error spike: "Cannot read properties of undefined (reading 'stripeId')"
14:03 UTC — error_rate jumps to 28%, latency spikes to 320ms
14:04 UTC — Pods begin CrashLoopBackOff, 7 restarts
14:05 UTC — error_rate: 91%, latency: 1450ms, API Gateway circuit breaker opensAn agent reasoning over get_recent_deployments → get_service_logs → get_service_metrics →
analyze_root_cause should produce a hypothesis pointing to v2.4.1 as the culprit with ~0.90 confidence.
Quick Start
# 1. Clone / navigate to project
cd incident-commander-mcp
# 2. Install dependencies (already done by CLI)
npm install
# 3. Start development server
npm run devBuilding for Production
npm run build
npm startBuilt with NitroStack ⚡
Available Tools
8 toolsanalyze_root_causeARead-onlyIdempotent
Performs automated root cause correlation for a service exhibiting a given symptom. Internally cross-references recent deployments, error logs, and service metrics to produce a structured hypothesis: {likely_cause, confidence (0–1), evidence[]}. This is the primary reasoning tool — call it after collecting data with get_recent_deployments, get_service_logs, and get_service_metrics. Confidence above 0.7 suggests high certainty.
Inputs: service (required), symptom (required).
check_pod_statusARead-onlyIdempotent
Checks the current Kubernetes pod/container status for a service. Returns desired vs ready replica counts, restart count, pod phase (Running, CrashLoopBackOff, etc.). A pod_phase of "CrashLoopBackOff" combined with high restarts is a strong indicator of a bad deploy. Use this alongside get_service_metrics to assess service health during an incident.
Inputs: service (required).
create_incident_ticketA
Creates a new incident ticket and appends it to the in-memory incident registry. Use this when a new issue is detected that warrants formal tracking. Returns the generated ticket ID and the full incident object. IMPORTANT: This tool appends a new record — calling it twice creates two tickets. Check for existing incidents with the incident://current resource before creating a new one.
Inputs: title (required), severity (required), description (required).
get_recent_deploymentsARead-onlyIdempotent
Retrieves the most recent deployments across all services, or filtered by a specific service name. Use this tool first during incident investigation to check whether any recent deployment correlates with an observed symptom. Each result includes service name, version, commit SHA, deployment time, and who triggered the deploy.
Inputs: limit, service.
get_service_logsARead-onlyIdempotent
Retrieves error logs for a specific service within a time window. Use this to identify error messages, stack traces, and patterns after a symptom is observed. Filter by log level to focus on ERROR/FATAL entries during an incident. The "since_minutes" parameter looks back N minutes from the current simulation time (14:06 UTC).
Inputs: service (required), level, since_minutes.
get_service_metricsARead-onlyIdempotent
Retrieves time-series performance metrics for a specific service: latency (ms), error rate (0–1), CPU usage (0–1), memory usage (0–1), and requests per second. Use this to identify performance degradation trends and correlate them with deployments or incidents. Look for sudden spikes in error_rate or latency_ms following a deployment.
Inputs: service (required), since_minutes.
post_status_updateA
Appends a timeline entry to an existing incident record. Use this to log investigation findings, actions taken, or status changes during an ongoing incident. Requires a valid incident_id (e.g., "INC-2026-0042"). The message will be timestamped and attributed to "incident-commander-mcp".
Inputs: message (required), incident_id (required).
trigger_rollbackADestructive
Submits a rollback request for a service to a target version. IMPORTANT: This tool does NOT immediately execute the rollback. It returns a pending_approval object with a rollback_id that must be reviewed and approved by a human operator before any deployment action is taken. Call analyze_root_cause first to confirm the target version is safe to roll back to. Required fields: service name and target_version (e.g., "v2.4.0").
Inputs: service (required), target_version (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 8 tools
Each tool targets a distinct data type or action: deployments, logs, metrics, pod health, root cause analysis, rollback, ticket creation, and status updates. There is no overlap in purpose, and the descriptions clarify the unique role of each tool.
All tool names follow a consistent verb_noun pattern with lowercase snake_case: get_recent_deployments, check_pod_status, analyze_root_cause, trigger_rollback, create_incident_ticket, post_status_update, etc. The naming convention is uniform and predictable.
With 8 tools, the server is well-scoped for incident management. It provides a balanced set of investigation, analysis, action, and documentation tools without unnecessary redundancy or overwhelming volume.
The tool set covers the core incident lifecycle: investigate (deployments, logs, metrics, pods), analyze (root cause), act (rollback), and document (create ticket, post updates). A minor gap is the lack of a tool to list or retrieve existing incident details, though the incident://current resource partially addresses this.
Maintenance
Related MCP Connectors
AI-powered incident management and server monitoring via MCP.
Read-only MCP server for AIStatusDashboard status, incidents, metrics, and fallback recommendations.
MCP server for Uptrack uptime monitoring. Manage monitors and incidents from AI agents.
MCP-native AI SRE: ask what's broken in production, get a reviewed GitHub fix PR.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceA multi-agent MCP server that turns LLMs into an autonomous incident-response copilot, enabling rapid investigation, correlation, and remediation of production incidents.MIT
- AlicenseNot gradedqualityCmaintenanceAI-powered incident management and server monitoring via MCP.MIT
- FlicenseNot gradedqualityCmaintenanceAn AI-native incident response server that exposes diagnostic tools (system status, error logs, ticket creation) via MCP, enabling LLM agents to autonomously assess and respond to incidents.-
- FlicenseNot gradedqualityDmaintenanceAn AI Operations Investigation MCP server that enables LLMs to investigate order fulfillment incidents by correlating data across independent services using predefined scenarios.-