Results Watcher

[satand]

In my opinion the actual documentation (https://tekton.dev/docs/results/watcher/) is not clear about how to set the watcher completed_run_grace_period, check_owner, store_deadline, disable_storing_incomplete_runs flags in particular if you are using Tekton Operator and it TektonConfig resource.

[AlanGreene]

Thanks for raising this issue.

@tektoncd/results-maintainers fyi

[enarha]

Note taken. I'll check the Tekton Results configuration exposed through the TektonConfig and address the potential gaps.

[enarha]

@satand I started working on this. While most of the Tekton Results API configuration is available though the TektonConfig, none of the Watcher configuration is. Aside from the TektonConfig gap, another difference between the api and the watcher is that the former uses configmap for its configuration, while the latter uses arguments passed to the watcher container, usually through the watcher deployment. So I'm currently working on closing the gap with the TektonConfig. In the meantime, the best place to look for "documentation" of the watcher parameters is actually the code, where the parameters are defined: https://github.com/tektoncd/results/blob/main/cmd/watcher/main.go#L63-L84
I'll also update the watcher documentation (https://github.com/tektoncd/results/tree/main/docs/watcher#readme) to further document those parameters.