GitHub Actions Runners & Org Health

Monitor your organization's self-hosted GitHub Actions runners and overall repo health in one place. Built for github-actions-runner-exporter, a small open-source Go exporter (MIT).

What it shows

Runners

• Runners online, offline, and busy right now
• Per-runner online and busy timelines, so you can see capacity and queueing over the day
• Exporter health: whether the runner and org-stats polls are fresh or serving a stale cache

Organization

• Public and private repository counts
• GitHub API rate limit remaining, now and over time (lowest active budget window)
• Open pull requests, org-wide and per repository, with a click-through to the PRs on GitHub
• Dependabot alerts by severity, org-wide and per repository, with a click-through to each repo's alerts
• CI health table: the latest run of every active workflow per repo, with its conclusion and a link to the run

CI History

• Job run and queue time table for the selected time range, per repo, workflow and job
• Average and p95 job run time per job
• Queue time (average and p95) by runner pool, i.e. per runs-on label, for example x64 vs arm64
• Queue time by individual runner, with GitHub-hosted runners grouped together
• Completed workflow runs by conclusion (success, failure, cancelled and so on) over time

Variables

• Prometheus data source: select your instance on import; nothing is hardcoded to a specific one
• GitHub org and Repo: filter the panels
• Rate window: the rate window (6h, 1d or 7d) used by the CI History panels

Requirements

• Prometheus
• github-actions-runner-exporter, with a GitHub token that can read your organization's self-hosted runners, plus repository read access to Pull requests, Actions and Dependabot alerts for the Organization and CI History panels (see the exporter's README for details)

Getting started

1. Run the exporter

   docker run -d -p 9222:9222 \
     -e GITHUB_ORG=your-org \
     -e GITHUB_TOKEN=your-token \
     ghcr.io/drumandbytes/github-actions-runner-exporter:latest
   

   On Unraid, it's also in Community Apps (search for github-actions-runner-exporter).

2. Scrape it from Prometheus

   scrape_configs:
     - job_name: github-actions-runner-exporter
       static_configs:
         - targets: ['exporter-host:9222']
   

3. Import this dashboard and select your Prometheus data source.

Notes

• Runner status is polled on a short cache (15s by default). Org and repo stats use a longer one (5 min) to stay well within GitHub's API rate limit.
• CI status is tracked per workflow, so a failing workflow can't hide behind a later, unrelated green run.
• Job queue and run times are recorded by the exporter as Prometheus histograms, so the CI History panels build up over time. Prometheus is the only store, so how far back they go depends on your retention. After the exporter starts, history fills in from that moment; earlier runs aren't backfilled.

Source, issues and the dashboard JSON: https://github.com/drumandbytes/github-actions-runner-exporter