Announcing Support for Kubernetes v1.35

We are pleased to announce that Avisi Cloud now supports Kubernetes v1.35, themed "Timbernetes". This version is available for both new and existing clusters via the Console. Check out what's coming next by testing these latest improvements on your staging or new clusters.

Please see our documentation for guidance on planning and performing your cluster upgrade.

If you are using update channels to manage your cluster versions, note that upgrading to a new minor Kubernetes version requires you to manually change the update channel to v1.35. We recommend setting up an upstream cluster so you only need to update the channel once. Downstream clusters will then follow automatically.

Key New Features

In-Place Pod Resize is now Stable After more than six years of development, In-Place Pod Resize reaches General Availability. You can now adjust CPU and memory resources of running containers without restarting Pods. This enables zero-downtime vertical scaling for latency-sensitive workloads, dynamic resource management during startup phases, and seamless integration with Vertical Pod Autoscaler (VPA). Memory limit decreases are now also permitted, and deferred resizes are retried based on PriorityClass when node capacity is insufficient.
PreferSameNode Traffic Distribution (Stable) The trafficDistribution field for Services gains a new PreferSameNode option that strictly prioritizes endpoints on the local node before falling back to remote endpoints. The existing PreferClose option has been renamed to PreferSameZone for clarity. This gives operators fine-grained control over traffic routing to reduce latency and cross-zone network costs.
Fine-grained Supplemental Groups Control (Stable) Pods can now precisely control which supplemental groups are applied to container processes. This closes a long-standing security gap where container images could inject unexpected group memberships, improving workload isolation and compliance posture.
Job ManagedBy field (Stable) The managedBy field for Jobs reaches GA, allowing external controllers to take ownership of Job lifecycle management. This is particularly useful for frameworks like Kueue that need to manage job queuing and scheduling independently from the built-in Job controller.

Platform Upgrades

With Kubernetes v1.35, our managed images ship significant platform-level upgrades:

Ubuntu 24.04 LTS — Our v1.35 images run on Ubuntu 24.04 LTS (Noble Numbat), upgraded from Ubuntu 22.04 LTS. This brings a newer kernel, improved security defaults, and extended support until 2029.
Calico v3.31.4 — Upgraded from v3.30 with improved network policy performance and stability.
Cilium v1.19.1 — Upgraded from v1.18, the default CNI for new clusters, bringing performance improvements and new networking features.
containerd 2.2.2 & runc v1.4.1 — Upgraded from containerd 2.1.x and runc 1.3.x.

Available Versions

v1.35.3-u-ame.4

Useful Links

Release notes

Please see our release notes for the full changelog, including validated component versions in our managed images (Kubernetes v1.35.3, Cilium v1.19.1, CNI plugins v1.9.1, containerd 2.2.2, runc v1.4.1).