Kubernetes 1.35: A Deep Dive into “Timbernetes”

The Cloud Native Computing Foundation (CNCF) has unveiled the latest release of Kubernetes, version 1.35, aptly named “Timbernetes.” This release shines a spotlight on mutability and the optimization of high-performance AI and machine learning workloads, making it a significant milestone for developers and operators alike.

In-Place Pod Resize: A Game Changer

Among the key features of Kubernetes 1.35 is the introduction of In-Place Pod Resize, now generally available. This innovative feature enables cluster operators to modify CPU and memory resources for running pods without the need to restart the container. This improvement is particularly beneficial as it streamlines resource management, reducing downtime and enhancing application performance.

Quote from Piotr Mińkowski: “Why is it essential for Java? You can give the pod extra CPU only for startup, then shrink it after. App starts fast, the Pod uses the right amount of resources always.”

Alpha Features: Gang Scheduling and PodGroup API

The Kubernetes 1.35 release also introduces exciting alpha features, notably, the native support for Gang Scheduling within the scheduler. This functionality is crucial for scheduling a group of related pods—particularly in AI/ML training scenarios—ensuring they are executed simultaneously or not at all. Coupled with this is the new PodGroup API resource, which allows users to define their scheduling requirements directly within the core API, streamlining the process significantly.

Enhanced Tooling with /flagz and /statusz

Further enhancing the usability of Kubernetes components, the release includes improvements to the /flagz and /statusz endpoints. These modifications provide machine-parsable output for authorized users, facilitating easier monitoring and automated troubleshooting. This makes observability tools more efficient by allowing them to interact with core components through HTTP without excessive text parsing.

Horizontal Pod Autoscaler and Certificate Management Improvements

Kubernetes 1.35 also brings significant upgrades to the Horizontal Pod Autoscaler (HPA), which now offers a configurable tolerance that has graduated to beta and is set as a default feature. In previous iterations, HPA relied on a fixed 10% tolerance for scaling actions. This version empowers users to define customized tolerance levels per resource, adapting better to diverse workload requirements without necessitating configuration changes across the cluster.

Notably, the PodCertificateRequests API objects have moved to beta status as well. This enhancement automates the process by which Kubernetes pods obtain certificates, managing certificate generation directly at the pod level. This simplification is key for implementing mutual TLS (mTLS) without the complications of bearer tokens or manual interactions.

Goodbye to Ingress NGINX

While not incorporated into version 1.35, the Kubernetes community has made a significant decision to sunset the Ingress NGINX controller, signaling a shift towards more integrated solutions. Users are encouraged to explore modern alternatives like the Gateway API, which focuses on both L4 and L7 routing capabilities. Ingress NGINX will only receive best-effort maintenance until March 2026, with a strong recommendation to migrate to the Gateway API or adopt an alternative third-party ingress controller.

A Comprehensive Release

Kubernetes version 1.35 boasts an impressive catalog of over 60 enhancements, including 22 alpha features, 19 transitioning to beta, 17 achieving general availability or stability, along with some deprecations and removals. This comprehensive release highlights the Kubernetes community’s commitment to continuous improvement and innovation.

Get Involved

For a more detailed look at the Kubernetes 1.35 release, developers and operators should refer to the official release notes and documentation. These resources offer an in-depth overview of enhancements, as well as insights into deprecations and removal of features.

Additionally, users are invited to join the release team for a live webinar on January 14, 2026, at 11:00 AM (CST). This event promises to be a valuable opportunity for gaining insights and asking questions about the latest release.

Looking to the future, Kubernetes version 1.36 is anticipated to arrive by April 2026, promising even more enhancements and features to support the evolving needs of developers and organizations.