Simulate in dev00
-
Already visible in dev00
-
Hypothesis:
-
If we need to adapt livez Probe configuration --> https://github.com/bettermarks/bm-backend/pull/2903
- default 1
- liveness 10
--- Things we tried so far ---
- Rule out NAT Gateway -> Logic says should not have any impact
- We enable all IPs to mongodb -> 0.0.0.0
- We manually routing rules from existing nodes.
- We delete daemonset.
- If CPU limits has any impact:
- Can verify by adding limits.
- Changed https://uwsgi-docs.readthedocs.io/en/latest/Options.html#worker-reload-mercy to 600 seconds for dev00 -> https://github.com/bettermarks/bm-backend/pull/2902
- terminationGracePeriodSeconds -> Also ruled out as logs doesn't make sense either. https://github.com/bettermarks/bm-backend/pull/2900/files -> Definitely not helping and it is not the case
- We tried changing uwsgi behaviour but it has no impact on the currently observed behaviour of https://github.com/bettermarks/bm-backend/pull/2898/files -> Definitely not helping and it is not the case
- We also changed startup Probe -> It is an old container so should not be a problem
- New one is probed too early and somehow it causes problem with old pods
Change Log:
- change deployment strategy -> Should not have any impact
- IONOS changed something -> livez is not answering when deployment is happening.