GitHub Actions Runner Controller 0.15.0 Cuts API Traffic for Large Runner Fleets

Operators running big self-hosted runner pools get tunable rate limits, concurrency and shutdown timing, plus lighter load on the Kubernetes API.

GitHub has shipped Actions Runner Controller 0.15.0, a Kubernetes-based autoscaling system for self-hosted GitHub Actions runners, and the update is aimed at making large runner fleets cheaper to operate and easier to run reliably. The release switches controller updates to Kubernetes patch requests, moves runner status aggregation into metrics for EphemeralRunnerSet and AutoscalingRunnerSet, and filters events so the controller does less unnecessary reconciliation. It also adds knobs for listener API rate limits and reconciliation concurrency, which lets operators tune throughput against API pressure in bigger deployments. On the lifecycle side, patch upgrades now preserve resources in place, shutdown timing is configurable, the controller can recover when a runner scale set is missing from GitHub’s service record, and ephemeral runners can be removed faster after successful pod exit.