Kubernetes: Gang Scheduling, Watch Cache Hardening, and DRA Fixes

Workload-aware scheduling and watch infrastructure dominated activity, with gang Job lifecycle fixes and deeper pod group hierarchies, plus a push to harden watch cache and graduate Watch List. DRA device allocation correctness fixes round out the highest-impact changes.

Duration: PT2M46S

Episode overview

This episode is a short developer briefing from Kubernetes.

It explains recent repository work in plain language.

  • Show: Kubernetes
  • Published: 2026-11-16T14:03:35Z
  • Audio duration: PT2M46S

Transcript excerpt

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

Hello, it's Monday, November 16th, 2026, and this is your Kubernetes developer briefing.

The lead signal today is maturity at scale — the scheduler is getting cleaner gang semantics and deeper hierarchies, while the watch path is getting reliability fixes aimed at large clusters.

First, workload-aware scheduling. Merged pull request 143537 changes suspended gang Jobs to delete the pod group they control and recreate it on resume, instead of reusing stale state, with an event emitted on delete. That pairs with pull request 144193, which proposes raising the maximum pod group hierarchy depth…

Second, watch cache and Watch List hardening. Merged pull request 143958 resyncs watchers as they catch up, and pull request 143508 adds an override for the watch cache event limit. Around that, pull request 144195 proposes promoting Watch List and its client to general availability in 1.38, pull request 144201…

Third, device allocation correctness. Cherry picks in pull requests 144189 through 144191 fix a regression where admin access marked devices as in use and blocked later claims in the same pod, while pull requests 144200 and 144198 fix tracking when informer updates are skipped and prevent…

A…

Nearby episodes from Kubernetes

  1. Safer In-Place Resize and Scheduler Correctness
  2. Pod-Level Resources Go GA
  3. Scheduler Hardening and Pod Resource Fixes
  4. DRA Beta Push, Kubelet Restart Fixes, and Storage Compression
  5. Weekly Recap - Dynamic Pods, Device Sharing, and Writable Cgroups
  6. Watch Cache Hardening and Flake Cleanup
  7. Dynamic Pods and Device Scheduling Hardening
  8. DRA Admin Access Fixes and API Server Efficiency