Using Prometheus to Monitor GitLab Runner and GitLab CI Pipelines

GitLab Runner exposes Prometheus metrics via an embedded HTTP server at the /metrics endpoint. By enabling this endpoint—either by setting listen_address in config.toml or by adding the --listen-address flag when starting the runner—you can let Prometheus scrape runner‑specific metrics such as job counts, Go runtime statistics, and general system metrics.

Example config.toml configuration:

listen_address = "[::]:9252"
concurrent = 10
check_interval = 30
log_level = "info"

After restarting the runner, verify the listening port with:

bash-5.0$ netstat -anlpt | grep 9252

When port 9252 is LISTEN, the metrics endpoint is reachable. Retrieve a sample of the metrics with:

curl 127.0.0.1:9252/metrics

Next, add a job to the Prometheus configuration to scrape the runner metrics:

- job_name: 'gitlab-runner'
  metrics_path: '/metrics'
  scheme: http
  bearer_token: bearer_token
  static_configs:
    - targets: ['192.168.1.200:30092']

Visit http://192.168.1.200:30003/new/targets to confirm the target is up, then import a Grafana dashboard (e.g., dashboard 9631 ) to visualize the runner metrics.

1.2 Configuring GitLab CI Pipeline Monitoring

To obtain a unified view of CI/CD pipeline status across projects, deploy the gitlab-ci-pipelines-exporter . Clone the exporter repository, edit values.yaml to set the GitLab URL and token, and install the Helm chart:

git clone https://github.com/mvisonneau/gitlab-ci-pipelines-exporter.git
vim chart/values.yaml
# edit the gitlab section with URL and token
helm install gitlabci-pipeline-exporter --namespace gitlab-runner ./chart

Extend the Prometheus configuration with a new job for the exporter:

- job_name: 'gitlab-runner-ci-pipeline'
  metrics_path: '/metrics'
  scheme: http
  bearer_token: bearer_token
  static_configs:
    - targets: ['10.1.234.132:80']

Finally, import a Grafana dashboard for CI pipeline metrics (e.g., dashboard 10620 ) to display pipeline health alongside runner statistics.