Engagement Picks
Curated weekly reads with a short, opinionated “why I care”. DevOps / SRE / MLOps / AIOps, high-signal, low-noise.
Why I care: The in-place pod resize feature finally stabilizes. If you run HPA + VPA together you'll want to read the interaction notes carefully before upgrading.
Why I care: Alert state history now writes to a dedicated time-series store. This changes how you think about alert flap detection and long-term trend queries.
Why I care: Stacks aren't just 'workspaces with better DX'. The dependency graph changes how you structure cross-environment rollouts. Worth the 20-minute read.
Why I care: The multi-cluster adoption number jumped 12 points. The failure modes teams report (networking complexity, upgrade fear) match exactly what I see on the ground.
Why I care: EKS Auto Mode is the headline but the quiet win is improved IRSA with pod identity. Much cleaner permission scoping for multi-tenant clusters.
Why I care: Native histograms reduce cardinality by 60–80% on high-label metrics. The migration path is non-trivial but the storage savings are real.
Why I care: The init container lifecycle and sidecar renewal loop are frequently misunderstood. This clears up why secrets go stale on long-running pods.
Why I care: ZooKeeper removal changes your ops runbook significantly. Broker discovery, partition reassignment, and rolling restart procedures all need updating.