Tekton Pipelines v1.13.0: Compressed Results, Timeout Fixes, and Observability

Tekton Pipelines v1.13.0 brings termination message compression to fit more results in the 4KB limit, fixes finally tasks not running on timeout, and adds OpenTelemetry tracing to the notifications controller.
By Vincent Demeester, Red Hat |

We’re excited to announce the release of Tekton Pipelines v1.13.0 “Pixie-bob Project 2501”!

🎉 Squeezing more out of every pipeline: compressed results & timeout fixes 🎉

Termination Message Compression

The biggest new feature is optional termination message compression (#9682). Kubernetes limits termination messages to 4KB, which constrains how many results a Task step can produce. With the new alpha feature flag enable-termination-message-compression, Tekton now uses flate compression to fit approximately 5.7x more results in the same 4KB limit.

Key details:

  • Zero new dependencies — uses Go stdlib only
  • Full backward compatibility — the parser auto-detects compressed messages
  • Enable with: enable-termination-message-compression: "true" in feature-flags ConfigMap (requires enable-api-fields: "alpha")

Finally Tasks Now Run After Tasks Timeout

A long-standing bug (#6794) caused finally tasks to be skipped when the tasks timeout was exceeded. The PipelineRun was immediately marked as Failed without giving finally tasks a chance to run — defeating the purpose of having cleanup tasks.

This is now fixed (#9709). When tasks timeout but finally tasks still need to run, the pipeline continues with reason PipelineRunTimeoutRunningFinally until finally tasks complete, mirroring the existing CancelledRunningFinally pattern.

TEP-0137: CloudEvents Formats Configuration

Building on v1.12.0’s notifications controllers, the formats field in config-events is now active (#9776). The default value is tektonv1, which preserves existing behaviour. Setting an invalid or unrecognised format value logs a warning and suppresses event emission for that format.

OpenTelemetry Tracing for Notifications

The TaskRun notifications controller now emits OpenTelemetry spans (#9912), covering ReconcileKind, ReconcileRunObject, and EmitCloudEvents. This enables operators to trace CloudEvent delivery latency end-to-end.

Additional Fixes

  • Affinity assistant volume names: Long workspace names no longer cause StatefulSet failures — names are automatically truncated with a hash suffix to stay within the 63-character Kubernetes limit (#9752)
  • Default resource requirements: Internal containers (entrypoint, sidecar-logresults, etc.) now inherit default resource requirements from config-defaults, preventing scheduling issues on clusters with LimitRange policies (#9366)
  • Resolver cache TTL: Per-resolver TTL values in individual resolver ConfigMaps are now correctly respected instead of always falling back to the global TTL (#9625)
  • Resolver validation: Resolvers now only permit resolving StepActions, Tasks, and Pipelines — arbitrary data resolution is no longer allowed (#9588)

Upgrade Notice

Resolver validation (#9588): Tekton Resolvers are now only permitted to resolve StepActions, Tasks, and Pipelines. Custom resolvers or ResolutionRequests which use the Resolver API for other object types will no longer function.

Get Started

Install or upgrade to v1.13.0:

kubectl apply -f https://infra.tekton.dev/tekton-releases/pipeline/previous/v1.13.0/release.yaml

Check out the full release notes and documentation for more details.

Have questions or feedback? Join us on Tekton Slack or open an issue on GitHub.