Automating Kubernetes Upgrades for Zero Downtime
summary
Claims that upgrades run with zero downtime after building an automated pipeline that uses the cloud provider’s automated upgrade feature and release channels, plus pre-pro and staging gates. The team converts stateful workloads to persistent volumes, standardizes rolling updates, readiness/liveness probes, PDBs, HPA and graceful shutdown, and verifies on five clusters totaling 10–12,000 pods. Manual upgrades once take weeks; the automated process runs on a schedule (dev weekly, staging bi‑monthly, prod every four months) and saves operational time.
AI-written summary · about
chapters (11)
- 0:00Kubernetes upgrades intro
- 1:03Why upgrading Kubernetes is challenging
- 3:51Reliable upgrade foundation
- 10:15Key automation considerations
- 13:03Upgrade methods
- 15:31Kubernetes release cycles
- 18:19Choosing upgrade channel
- 20:45Upgrade timeline design
- 22:29Automation benefits
- 23:55Automation trade-offs
- 24:16Final recommendations
related talks
37mZero downtime Deployment on Kubernetes
18mAutomating Kubernetes Cluster Updates: Achieving Zero Downtime Effortlessly
35mZero-Downtime Deployment on Kubernetes
40mStrategies for Zero Down Time, Frequent Deployments
46mEnd-Users Won't Even Notice, Painless Kubernetes Upgrades
23mHow To Migrate Kubernetes Cluster With Zero Downtime
53mDeploying to Kubernetes Thousands of Times Per/Day
10mGenesis Startup Competition Pitch