Kubernetes: Dynamic Pods and Device Scheduling Hardening
Kubelet and scheduler work converged on making running pods and assigned devices easier to change safely, with dynamic container adds and DRA allocation fixes leading the day alongside controller recovery work.
Duration: PT2M38S
Episode overview
This episode is a short developer briefing from Kubernetes.
It explains recent repository work in plain language.
- Show: Kubernetes
- Published: 2026-11-14T14:02:11Z
- Audio duration: PT2M38S
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 is Saturday, November 14th, 2026.
The through-line today is mutability: Kubernetes is making running pods and assigned devices easier to change safely, while hardening the controllers that manage them.
First, dynamic containers. Pull requests 143714 and 143715 add kubelet and API support for adding and removing containers without recreating the pod. Adds and removals with resources go through resize allocation and admission, image updates are gated on allocation, and any blocked change surfaces as pod resize…
Second, device scheduling is getting more reliable and faster. Pull request 144115 fixes D-R-A admin access so an admin claim no longer marks a device as in use for others in the same allocate call. Pull request 144080 fixes a cross-wiring in the dynamic resources scheduler plugin where pending and fresh results…
Finally, control plane robustness. Pull requests 143717 and 143766 wire in leader election recovery and a pause during recovery, alongside a new controller robustness test framework in 139657. Conditional authorization adds C-E-L evaluation in 143024, and the scheduler now requeues an updated pod once for all its…
Smaller reliability fixes also landed:…
Nearby episodes from Kubernetes
- DRA Admin Access Fixes and API Server Efficiency
- DRA Preemption and Kubelet Reliability Push
- Stricter Pod Admission and Watch Cache Reliability
- Quantity Math Fixes and Watch Cache Hardening
- Storage Safety, API Scale and Scheduler Fixes
- Weekly Recap - Cache Scale-Up and Workload Hardening
- Fixing Stuck States and Tightening Node Behavior
- API Cache Overhaul and Clearer Job Status