Actions Runner Controller 0.15.0 adds configurable concurrency and rate limits
GitHub says Actions Runner Controller 0.15.0 adds configurable controller concurrency and listener client rate limits, alongside changes intended to reduce upgrade disruption and Kubernetes API requests.
SessionWatcher editorial · Published · Updated · Source announcement: 2026-10-01
Operational changes in 0.15.0
GitHub says Actions Runner Controller 0.15.0 includes reliability, scalability, and observability improvements for runner scale sets. Patch version upgrades now update resources in place, which GitHub says reduces disruption between autoscaling and ephemeral runner sets.
For operators tuning throughput, controller concurrency can be configured globally and per controller with max-concurrent-reconciles flags. Listener Kubernetes client rate limits are configurable through QPS and burst settings.
Sources: Actions Runner Controller release 0.15.0 - GitHub Changelog
Fewer disruptions and API requests
The release also makes controller shutdown behavior configurable through terminationGracePeriodSeconds, aligned with the controller manager graceful shutdown timeout. A scale set can be reregistered if its recorded counterpart no longer exists in the Actions service.
The controller now uses patch requests instead of full update requests. Runner status aggregation moved to metrics for EphemeralRunnerSet and AutoscalingRunnerSet, reducing status patch requests. GitHub also says incoming events are filtered to reduce reconciliations, and ephemeral runners are deleted faster when their pod exits successfully.
Sources: Actions Runner Controller release 0.15.0 - GitHub Changelog
AI assisted reporting, checked against the linked official sources. Source pages checked 2026-10-01. Editorial process and corrections.