Kubernetes: Watch Cache Scaling and Quantity Hardening
The API server watch path dominated activity with performance and correctness work, alongside a focused cleanup of resource quantity math. Smaller workload and documentation fixes round out a reliability-heavy day.
Duration: PT2M26S
Episode overview
This episode is a short developer briefing from Kubernetes.
It explains recent repository work in plain language.
- Show: Kubernetes
- Published: 2026-10-16T13:03:28Z
- Audio duration: PT2M26S
Transcript excerpt
This excerpt keeps the crawler page concise. Listen to the episode or use the RSS feed for the full update.
Good morning, it's Friday, October 16th, 2026.
The lead story is scale in the API server watch path, with a parallel push to make resource quantity math safer and more predictable.
First, the watch cache. A merged change in PR 138307 switches initial watch events to indexed, prefix-scoped snapshot lookups shared with the list path, aimed at large clusters — testing cited 150 thousand pods across thousands of namespaces. That pairs with PR 142760, which adds sharded list and watch helpers,…
Second, quantity handling. Three connected cleanups from the same effort limit precision on add and subtract in PR 143075, make scale conversion consistent between integer and decimal backends in PR 142817, and in open PRs 143085 and 143084 avoid expanding large exponents and keep scale away from the minimum 32-bit…
Finally, workload correctness and docs. PR 142902 fixes Deployment rollouts failing to report timeouts after old Replica Set scale-down errors. On documentation, PR 142541 corrects Stateful Set max unavailable to reflect round-down behavior, and PR 143099 flags the Endpoint Slice ports doc still saying 100 instead…
What's next: watch for reviews on the sharded informer rollout, the…
Nearby episodes from Kubernetes
- DRA Stability and Scheduler Backports
- Node Env Fixes and Control-Plane Reliability
- Gang Scheduling Speedups and API Server Efficiency
- Scheduler Hardening, Job Counting Fixes, and kubectl WebSockets
- Watch Costs and Pod Group Preemption
- Weekly Recap - Go 1.27 Hardening, Scheduling and Storage Correctness
- Storage Efficiency and Scheduler Resilience
- API Stabilization and Controller Reliability Fixes