Open
Description
Description
We're currently upgrading the apmsoak deployment manually during the feature freeze cycle. Ideally, every time a new version is available in ESS, an upgrade will be triggered . This will prevent us from forgetting to upgrade the deployment and having an outdated version running for too long
Challenges
- Version health, sometimes BCs aren't healthy, we should not upgrade if BC aren't healthy. An option to avoid that is to trigger the soak test upgrades (or check) only after the smoke tests pass.
- We need to have yet another job that is triggered with some periodicity, perhaps having a downstream job from the Smoke tests is the best option.
- It may be necessary to either use
docker_image
override to trigger an update when the underlying docker image changes, but not the version, or issue a cluster restart call, which can act the same way. The former is preferred to the latter.