-
Notifications
You must be signed in to change notification settings - Fork 77
Description
Problem Statement
Assert step can take a while, it can get interrupted / terminated, which can lead to stale resources in the cluster. I appreciate this is less of an issue on purely ephemeral clusters, but more of an issue on non throwaway clusters.
Solution Description
On interrupt / term and similar, clean up the resources created. It can be handled by a flag if we think keeping resources as they are in the cluster is useful for debugging. There is also no indication that no cleanup happened, so not sure if user should be warned of this even if we do no cleanup.
Alternatives
We can preseed a namespace and bind resource creations in there in chainsaw, then ensure it gets fully deleted outside of chainsaw, which should work fine for non global (namespace scoped) resources.
Additional Context
No response
Slack discussion
No response
Research
- I have searched other issues in this repository and mine is not recorded.