Prometheus MCP
Prometheus PromQL query MCP with both instant and range queries, metric discovery, and target inspection. 429 GitHub stars and 9 commits on main in the last 30 days. Self-hosted, runs against any Prometheus endpoint.
“Prometheus PromQL query MCP with both instant and range queries, metric discovery, and target inspection. 429 GitHub stars and 9 commits on main in the last 30 days. Self-hosted, runs against any Prometheus endpoint.”
INSTALL THIS SERVER
{
"mcpServers": {
"prometheus": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"PROMETHEUS_URL",
"ghcr.io/pab1it0/prometheus-mcp-server:latest"
],
"env": {
"PROMETHEUS_URL": "http://your-prometheus:9090"
}
}
}
}
{
"mcpServers": {
"prometheus": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"PROMETHEUS_URL",
"ghcr.io/pab1it0/prometheus-mcp-server:latest"
],
"env": {
"PROMETHEUS_URL": "http://your-prometheus:9090"
}
}
}
}
{
"mcpServers": {
"prometheus": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"PROMETHEUS_URL",
"ghcr.io/pab1it0/prometheus-mcp-server:latest"
],
"env": {
"PROMETHEUS_URL": "http://your-prometheus:9090"
}
}
}
}
{
"mcpServers": {
"prometheus": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"PROMETHEUS_URL",
"ghcr.io/pab1it0/prometheus-mcp-server:latest"
],
"env": {
"PROMETHEUS_URL": "http://your-prometheus:9090"
}
}
}
}
{
"mcpServers": {
"prometheus": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"PROMETHEUS_URL",
"ghcr.io/pab1it0/prometheus-mcp-server:latest"
],
"env": {
"PROMETHEUS_URL": "http://your-prometheus:9090"
}
}
}
}
6 TOOLS AVAILABLE
OUR ASSESSMENT
- 429 GitHub stars and MIT licence.
- 9 commits on main in the last 30 days.
- Six-tool surface covers query, discovery, and targets.
- Docker image at ghcr.io/pab1it0/prometheus-mcp-server for container-native deployment.
- PromQL passthrough: full Prometheus query syntax available to agents.
- Tool list is configurable via environment variables (avoid context-window bloat).
- Health check tool useful for orchestration.
- Self-hosted: operator runs the container against their Prometheus endpoint.
- Read-only by design; alert-rule and target configuration live elsewhere.
- Authentication to Prometheus depends on the Prometheus deployment (basic auth, mTLS, or open).
The MCP forwards PromQL queries to the configured PROMETHEUS_URL. Scope the network path to the agent host: Prometheus has no built-in auth in default deployments. Use a reverse proxy or Prometheus Federation to scope read access by metric label or namespace.
Teams running Prometheus for metrics who want agents to query PromQL inside the editor; SRE workflows where the agent investigates a metric anomaly via list_metrics then execute_range_query; container orchestration that benefits from a health_check tool for the MCP itself.
TECHNICAL DETAILS
ADOPTION METRICS
// Reading this429 stars on the pab1it0/prometheus-mcp-server repo. 9 commits on main in the last 30 days.
// Reading thisSecond-ranked in monitoring; community-maintained with strong cadence and a six-tool surface.
SOURCES & VERIFICATION
We don't take any single directory's word for it. Before scoring, we cross-reference 4 public MCP sources, install the server ourselves against the clients we cover, and record when we last re-verified.
The same server, 4 different lenses. We reconcile these signals into our editorial score, which is why our number sometimes diverges from a directory-aggregate star count.
| Source | Their rating | Their star count | Their downloads | Last synced |
|---|---|---|---|---|
| AutomationSwitch This page | 4.2editorial | 429 | — | MAY 3, 2026 |
| PulseMCP | — unrated | unavailable | unavailable | MAY 3, 2026 |
| MCP.so | — unrated | unavailable | unavailable | MAY 3, 2026 |
| Glama | — unrated | unavailable | unavailable | MAY 3, 2026 |
| Smithery | — unrated | unavailable | unavailable | MAY 3, 2026 |
// Counts are directory-reported; we don't adjust them. Discrepancies usually come from different snapshot times or star-caching.
OTHER MONITORING MCP SERVERS
Grafana MCP Server
Vendor-official Grafana MCP server from Grafana Labs. Covers dashboards (search, get, summary, JSONPath extraction, patch), datasources (Prometheus, Loki, ClickHouse, CloudWatch, Elasticsearch), and direct query execution (PromQL, LogQL, InfluxQL/Flux, ClickHouse SQL).
Langfuse MCP
MCP server for Langfuse. Query LLM trace data, sessions, and observations from inside agents for debugging and observability. 85 GitHub stars and 16 commits on main in the last 30 days.
Axiom MCP
Official Axiom MCP server. Query Axiom datasets via APL (Axiom Processing Language) and inspect dashboards, monitors, and notifiers. 12 stars and 30 commits on main in the last 30 days.
Rollbar MCP
Official Rollbar MCP server: error tracking, occurrence triage, deploy tracking, and item lifecycle for AI agents. 25 stars and 13 commits on main in the last 30 days.
PagerDuty MCP
Official PagerDuty MCP server for incident management, on-call schedules, services, escalation policies, and event API integration. 66 GitHub stars and 5 commits on main in the last 30 days. Apache-2.0.
Datadog MCP
Official Datadog managed MCP server covering logs, metrics, traces, incidents, and dashboards. HTTP-only transport, hosted by Datadog with regional endpoints (US1, EU, etc.). 6 commits on main in the last 30 days. Stars are modest (35) reflecting recent launch.
DISCUSS YOUR
MCP REQUIREMENTS.
Evaluating a server, scoping an internal deployment, or working out whether MCP is the right fit at all. Start the conversation and we will point you at the right piece of the ecosystem.