Skip to main content
The Prometheus integration gives sureops agents direct access to your Prometheus instance for PromQL queries — used during diagnosis to surface error rate trends, latency spikes, and service health metrics, and during verification to confirm that metrics have returned to baseline after a fix.
If your Prometheus is already configured as a Grafana datasource, you likely do not need this integration — the Grafana integration covers Prometheus queries through a single MCP server. Connect Prometheus directly only if you run Prometheus standalone without Grafana, or if your Prometheus is not reachable through your Grafana instance.

What this integration enables

  • PromQL queries — agents run arbitrary PromQL against your Prometheus for metric diagnosis and verification
  • Service health checks — agents evaluate up, kube_pod_status_ready, kube_deployment_replicas_available, and similar series during incident investigation
  • Alert correlation — agents can query the metrics that triggered an alert to understand trend context

Prerequisites

  • An environment selected in the top-right context switcher — Settings → Environment → Integrations is scoped to that environment and stays disabled until one is selected
  • A running Prometheus instance reachable from your network
  • The prometheus-mcp server running and reachable from sureops
  • Optional: bearer token or basic auth credentials if your Prometheus instance requires authentication
Integration connections are per-environment. Connecting Prometheus here only applies to the environment that was selected in the context switcher at the time — it does not carry over to your other environments. Repeat this connect flow for each environment that needs Prometheus.

Setup

1

Stand up prometheus-mcp

Deploy an MCP server that speaks PromQL against your Prometheus instance. If you’re using the sureops demo environment, prometheus-mcp is already provisioned in-cluster — you can skip this step.For your own environment, deploy your MCP server of choice, pointed at your Prometheus URL with whatever auth (none, basic, or bearer) your Prometheus instance requires configured as its upstream. These Prometheus-side credentials configure your MCP server directly — they aren’t entered into sureops.
2

Connect in sureops

In sureops, go to Settings → Environment → Integrations and click Connect next to Prometheus.Fill in:
  • MCP Server Endpoint — the URL of your running prometheus-mcp server
  • Authentication Method — Bearer Token, API Key, or Basic Auth, matching how your prometheus-mcp deployment expects inbound requests to be authenticated
  • Token — labeled Bearer Token, API Key, or Password depending on the method you selected above. The form requires a value here even if your prometheus-mcp deployment doesn’t enforce its own auth — enter any placeholder value in that case
Click Save & Connect.
3

Optionally record the upstream auth method (after connecting)

Once connected, reopen the Prometheus integration from Settings → Environment → Integrations. A Settings section appears below the connection status with an Auth Method field (none, basic, or bearer) — a free-text note describing how your prometheus-mcp server authenticates to Prometheus itself. It’s informational only and configured after connecting, not part of the initial connect form.

Required credentials

Fields collected when you connect the integration: Fields configured afterward, via the integration’s Settings section once connected: Your Prometheus instance’s own URL, bearer token, or basic-auth credentials are never entered into sureops directly — they configure prometheus-mcp, which handles authentication to Prometheus on sureops’ behalf.

Common errors


Official documentation