Migrate triage-ops to a helm chart instead of kubectl commands
Context
Corrective Action of #1352 (closed)
Goal
I think it would be helpful to switch to helm charts for triage-ops. Doing so would help us to:
- Track the old/new versions of pods that are being deployed, and fail the deployment in case the pods are not up and running.
- Fail the GitLab deployment in case the helm upgrade fails, and the person that merged the MR would at least be notified.
- Be able to lint the charts to find issues with it early on (maybe the missing secret could have been spotted: not 100% sure about that)