Kubernetes: Weekly Recap - WebSockets Go Default as Scheduler and Cache Harden

50 pull requests and 17 commits centered on graduation and hardening. Kubectl streaming switched to WebSockets by default while sharded watches, Job and PodGroup fixes, and quantity cleanup advanced reliability.

Duration: PT4M11S

Episode overview

This episode is a short developer briefing from Kubernetes.

It explains recent repository work in plain language.

  • Show: Kubernetes
  • Published: 2026-10-19T09:12:51Z
  • Audio duration: PT4M11S

Transcript excerpt

This excerpt keeps the crawler page concise. Listen to the episode or use the RSS feed for the full update.

Good morning, this is your Kubernetes briefing for October 12th through October 19th.

50 pull request activity items, 17 additional commits this week.

The lead pattern is graduation and hardening. Remote command streaming moves to WebSockets by default, while scheduling, batch workloads, and control-plane caching get targeted correctness and performance fixes.

First, streaming. Three connected changes complete the shift away from SPDY under KEP forty oh six. Pull request 142861 makes kubectl use WebSockets first for exec, attach, copy and port-forward, with automatic fallback to the old protocol, and removes the temporary opt-in toggles. Pull request 142897 promotes the…

Second, scale. Pull request 142760 adds sharded list and watch helpers, plus shard selector options for shared informers and client generators. Pull request 142975 exposes shard information on unstructured lists to support that path. In parallel, the API server cache sees sustained investment. Pull request 142813…

Third, workload correctness and code health. Job controller fixes in 142922 and 142923 stop ready-count and creation-metric errors when pod deletes or creates fail, so automation sees accurate status.…

Nearby episodes from Kubernetes

  1. Scheduler Gang Cleanup, Proxy and User Namespace Backports
  2. Watch Cache Cleanup and Webhook Ordering
  3. Watch Cache Scale Push
  4. DRA Stability and Scheduler Backports
  5. Watch Cache Scaling and Quantity Hardening
  6. Node Env Fixes and Control-Plane Reliability
  7. Gang Scheduling Speedups and API Server Efficiency
  8. Scheduler Hardening, Job Counting Fixes, and kubectl WebSockets