Investigate issue with getting etcdadmcluster object due to unknown namespace
- Dominant language
- Go
- Stars
- 2.1k
- Forks
- 328
- Avg merge
- 1d 4h
- Merged PRs (30d)
- 9
Description
https://t.corp.amazon.com/P167063564
There are two issues called out:
1. The cache is not functioning properly, resulting in a null namespace. Logs have been attached for reference.
2. Cluster certificates were updated, but the etcdadmcontroller pod continued reporting them as invalid.
The log error message itself is clear and includes valid dates, yet the pod still flagged the certificates as invalid. This caused the kube-apiserver to become non-responsive for a period.
We need to figure out how to get rid of the specific errors from the first part and figure our why restarting the pods would cause the certificate to be counted.
Contributor guide
Research direction
Start with the attached logs and the linked incident report, then trace the etcdadmcontroller pod behavior around cache lookups and certificate validation. Investigate why the etcdadmcluster namespace becomes null and why updated certificates are considered invalid after pod restarts. Done means both error conditions are explained and the kube-apiserver remains responsive after the relevant restart and certificate-update scenarios.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- go
- Domain
- infrastructure, security
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100