Claude MarketplacesSubmit

Observability Mcp

thotischner/observability-mcp

If you're building AI agents that need to triage production incidents, this server normalizes queries across Prometheus, Loki, and other observability backends so your agent doesn't juggle multiple query languages. It exposes twelve read-only tools including topology graphs, blast radius analysis, and anomaly detection, all with server-side filtering to return summaries instead of raw metric floods. The author benchmarked a local 8B model going from 0/10 to 10/10 accuracy on cross-namespace blast radius questions when given the topology tools. Ships with a web UI for wiring up connectors, supports OIDC and OPA for multi-user deployments, and includes a chaos demo with autonomous incident response. Start it with npx, point it at your Prometheus endpoint, and wire it into Claude via HTTP transport.

GitHub stars: 7
tools: 8
transport: STDIO
env vars: 3
last updated: Nov 9, 2026

Tools

8 tools

Public tool metadata for what this MCP can expose to an agent.

list_sourcesDiscover configured backends and connection status

Discover configured backends and connection status

No parameter schema in public metadata yet.

list_servicesDiscover monitored services across all backends

Discover monitored services across all backends

No parameter schema in public metadata yet.

query_metricsQuery metrics with pre-computed summary stats

Query metrics with pre-computed summary stats

No parameter schema in public metadata yet.

query_logsQuery logs with error/warning counts and top patterns

Query logs with error/warning counts and top patterns

No parameter schema in public metadata yet.

get_service_healthHealth score combining metrics + logs (0–100)

Health score combining metrics + logs (0–100)

No parameter schema in public metadata yet.

detect_anomaliesCross-signal anomaly detection with robust (median/MAD + trend) analysis

Cross-signal anomaly detection with robust (median/MAD + trend) analysis

No parameter schema in public metadata yet.

get_topologyReturn the merged infrastructure graph (resources + edges) from every topology-capable connector, filterable by source/kind/scope

Return the merged infrastructure graph (resources + edges) from every topology-capable connector, filterable by source/kind/scope

No parameter schema in public metadata yet.

get_blast_radiusPivot on the universal RUNS_ON relation — "if this resource's host fails, who else fails?". Works for pod→node, vm→hypervisor, container→host

Pivot on the universal RUNS_ON relation — "if this resource's host fails, who else fails?". Works for pod→node, vm→hypervisor, container→host

No parameter schema in public metadata yet.

Sponsors

Promote your product hereReach 125k+ visitors monthly Promote your product hereReach 125k+ visitors monthly

Agent CaffeineSponsoredClose your lid, come back to finished work

SnugSponsoredThe clipboard manager for the AI era Promote your product hereReach 125k+ visitors monthly

WaxSponsoredBlock anything with supervision, without erasing your iPhone Promote your product hereReach 125k+ visitors monthly

knobsSponsoredSee your site on any phone, any day, in any language

Configuration

MCP_TRANSPORT

Set to 'stdio' when an MCP client spawns the package directly (the npx default starts the HTTP gateway + Web UI on :3000 instead).

PROMETHEUS_URL

Prometheus base URL to connect as a metrics source (optional; sources are also configurable via the Web UI / sources.yaml).

LOKI_URL

Loki base URL to connect as a logs source (optional).

Readme

observability-mcp

The unified observability gateway for AI agents.

One MCP server that connects to any observability backend through pluggable connectors, normalizes the data, adds robust anomaly analysis, and provides a web UI for configuration.

One MCP endpoint, every backend — so an agent triaging an incident asks one normalized question instead of juggling N vendor servers and their query languages.

0/10 → 10/10: the same 8B local model goes from hallucinating blast-radius answers to exactly correct ones once it gets this gateway's topology tools — measured, not asserted.

npx @thotischner/observability-mcp                                    # start (UI on :3000)
claude mcp add observability --transport http http://localhost:3000/mcp   # wire into Claude

Twelve read-only tools (readOnlyHint: true on every one) · server-side filter/aggregate so agents get numbers, not haystacks · For-Agents guide

View the full README on GitHub

Details

Repository: github.com/thotischner/observability-mcp
Category: Monitoring & Observability
Package: @thotischner/observability-mcp
Transport: STDIO
Registry: MCP registry, active

Related Monitoring & Observability MCP servers

1

Datadog Mcptantiope/datadog-mcp-serverFull Datadog API access: monitors, logs, metrics, traces, dashboards, and observability tools4stars

2

Datadogus-all/datadog-mcp-serverDatadog MCP — 165 tools for metrics, monitors, logs, APM, RUM, incidents, CI/CD, fleet3stars

3

Monitorbetterdb-inc/monitorBetterDB MCP server - Valkey and Redis observability for Claude Code and other MCP clients1.3Kstars

4

Mcp Agent Trace Inspectordbsectrainer/mcp-agent-trace-inspectorStep-by-step observability for MCP agent workflows0stars

5

Kibana Mcpmshegolev/kibana-mcpKibana/Elasticsearch MCP — log search, aggregations, index discovery, dashboards.0stars

6

Grafanagrafana/mcp-grafanaAn MCP server giving access to Grafana dashboards, data and more.3.7Kstars