When GitOps goes wrong, and other challenges for Strimzi platform teams
summary
A Helm values syntax bug in a specific Helm version produces empty output that removes all Kafka topic CRs and causes topic and data deletion, prompting addition of a mirrored failover cluster and changes to backup timing. Production clusters enforce disallow topic deletion in production via Kafka config or Kubernetes RBAC, certificates require exported monitoring and scheduled renewals, and upgrades follow staged strategies culminating in MirrorMaker2 DNS failover for blue‑green. Teams validate with canary clusters and automated smoke tests and prefer time‑tested tool versions.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:20What is GitOps
- 9:50Certificate issues
- 17:20Component life cycles
- 18:10Dependencies
- 19:07Different approaches
- 22:07Lessons learned
- 24:06Final advice
related talks
44mTaming Apache Kafka Using GitOps
34mHow to Survive Managing Strimzi in an Enterprise
36mDisaster Recovery in Action with Kafka and Strimzi
29mBeyond the Install: Take Full Ownership of Your Strimzi Cluster
42mTransition to Apache Kafka on Kubernetes With Strimzi
30mFrom Running to Operating: Real-World Strimzi in Production
27mThere's No Place Like Production
04mProject Lightning Talk: Strimzi – What's New and What's Next