GitSpider is GitHub Actions monitoring for small teams. It watches every workflow run, flags the slow, flaky, and wasteful ones, and comments on the pull request when its run fails or runs slow (and alerts Slack when a workflow stays slower), before your engineers feel it.
No signup. Try it on honojs/hono · vitejs/vite.
Three things. Done well.
When a pull request's CI run fails, GitSpider comments on it with the failing commit and a link to the run, and says when a run came in slower than the last week. The automatic slowdown notice waits until the slower level holds. The dev sees it where they already are, before it ships.
p95 duration, failure rate, time-to-recover, and throughput per workflow, with the bad PR flagged when today drifts past your baseline. Alerts debounce 6h so your channel stays sane.
26 detectors flag duplicated push/PR runs, missing concurrency & caches, always-failing jobs, silent cron death, Docker rebuilds and more, with the exact YAML fix, and recoverable minutes where they're honestly measurable.
Install → crawl → catch. Install in ~2 minutes, no YAML, no config.
Add the GitHub App to the repos you want watched: read access plus PR comments, no code writes. Two minutes, no YAML, no config.
GitSpider records every Actions run and learns each workflow's normal: duration, failure rate, billable minutes.
The moment a run drifts (slower, flakier, pricier), you're pinged in Slack, email, or right on the PR.
~2-minute install. Free to watch your CI. PR comments + Slack alerts on every paid plan.
Install on GitHub