Upgrading K8s: run as fast as we can, just to stay in place
summary
Claims that upgrades require upgrading everything and pinning every artifact. Recommends relying on Helm charts vetted on Artifact Hub, using Terraform as glue to pin infra and manifests, and keeping two debugging tools for redundancy. Notes a real failure from a missed nginx chart and frequent API or security changes that can break clusters. Advises annotating repetitive actions, keeping runbooks, and investing in observability: metrics, logs, kube events to detect unknown failures.
AI-written summary · about
related talks
19mUpgrading Your K8s Clusters: Key Considerations for a Smooth Transition
37mKubernetes Security and How to Fix K8s Cluster at Scale
42mMastering Kubernetes: Challenges, Strategies, and Best Practices
10mUsing Helm to the Fullest
25mAutomating Kubernetes Upgrades for Zero Downtime
43mKubernetes Operability Tooling
35mTo Infinite Scale and Beyond: Operating Kubernetes Past the Steady State
19mKube Me This! Kubernetes Ideas and Best Practices