Prometheus

Pull metrics and alerts from a Prometheus server using PromQL queries.

Edition: Community (CE)

Configuration

{
  "adapter": "prometheus",
  "name": "My Prometheus",
  "config": {
    "url": "http://prometheus:9090",
    "bearerToken": "optional-auth-token",
    "queries": [
      "up",
      "rate(http_requests_total[5m])",
      "node_memory_MemAvailable_bytes"
    ],
    "pullAlerts": true
  },
  "poll_interval": 30
}

Field	Required	Default	Description
url	Yes	—	Prometheus base URL
bearerToken	No	—	Bearer token for auth
basicAuth	No	—	{ username, password } for basic auth
queries	No	["up"]	PromQL queries to execute
pullAlerts	No	true	Also pull active alerts

What It Pulls

Metrics

Each PromQL query is executed via /api/v1/query. Results are normalized into metric objects with their Prometheus labels preserved.

Alerts

Active alerts fetched from /api/v1/alerts with alert name, severity, and firing state.

Prometheus Scraping Cockpit

Cockpit also exports metrics for Prometheus to scrape at /metrics:

scrape_configs:
  - job_name: 'cockpit'
    scrape_interval: 30s
    static_configs:
      - targets: ['cockpit:3000']
    metrics_path: '/metrics'
    authorization:
      type: Bearer
      credentials: your-metrics-token

This is a separate feature from the Prometheus integration — the integration pulls from Prometheus, while /metrics lets Prometheus pull from Cockpit.