Case Sensitive: Making Sure Input is User Friendly
summary
Cluster runs on Kubernetes 1.17 and uses the Horizontal Pod Autoscaler v2beta2; Helm chart templates and CI/CD set a default of values.yaml = "cpu" which inexplicably fails while "CPU" works, causing three days of debugging. Switching the HPA resource to memory succeeds, a PR to accept lowercase CPU goes stale after a year, and the concrete takeaway is to make inputs case-insensitive or ensure docs and accepted values exactly match.
AI-written summary · about
chapters (7)
- 0:00Start
- 0:43HPA Change
- 1:29Time To Verify!
- 1:56Time To Investigate!
- 2:58Let's try something...
- 3:26The Aftermath...
- 3:42Lessons Learned!
related talks
31mReworking Your Team for Kubernetes
32mAutomation and the Kubernetes Contributor Experience - Aaron Crickenberger, Samsung SDS
31mFrom Bottleneck To Breakthrough: Conquering Applications Star...
30mPhippy’s First Steps Into Kubernetes
21mBuilding High Quality OpenShift Applications
35mThe Oregon Trail to Kubernetes [I]
26mFive Hundred Twenty-five Thousand Six Hundred K8s CLI’s
01h 32mTutorial: Debug Your Kubernetes Apps