Description of problem:

Upstream issue - https://github.com/tektoncd/results/issues/514

We have encountered instances of race condition with aggressive pruning of PipelineRuns in Tekton Results Watcher while log streaming from the watcher to the API server which causes logs to be missing.

Prerequisites (if any, like setup, operators/versions):

Tekton Results with immediate pruning enabled.

Steps to Reproduce 

Upstream issue - https://github.com/tektoncd/results/issues/514

Actual results:

Missing logs

Expected results:

Logs should not be missing.

Reproducibility (Always/Intermittent/Only Once):

Build Details:

Additional info (Such as Logs, Screenshots, etc):

The probable cause is the pod gets deleted before watching could stream the logs to the API server from an aggressive pruning such as while `immediate pruning is activated` and that creates a race condition.