Kubernetes 1.32 is now available on GKE

Kubernetes 1.32 is now available in the Google Kubernetes Engine (GKE) Rapid Channel, just one week after the OSS release! For more information about the content of Kubernetes 1.32, read the official Kubernetes 1.32 Release Notes and the specific GKE 1.32 Release Notes.

This release consists of 44 enhancements. Of those enhancements, 13 have graduated to Stable, 12 are entering Beta, and 19 have graduated to Alpha.

Kubernetes 1.32: Key Features

Dynamic Resource Allocation graduated to beta

• Dynamic Resource Allocation graduated to beta, enabling advanced selection, configuration, scheduling and sharing of accelerators and other devices. As a beta API, using it in GKE clusters requires opt-in. You must also deploy a DRA-compatible kubelet plugin for your devices and use the DRA API instead of the traditional extended resource API used for the existing Device Plugin.

Support for more efficient API streaming

• The Streaming lists operation has graduated to beta and is enabled by default; the new operation supplies the initial list needed by the list + watch data access pattern over a watch stream and improves kube-apiserver stability and resource usage by enabling informers to receive a continuous data stream. See k8s blog for more information.

Recovery from volume expansion failure

• Support for recovery from volume expansion failure graduated to beta and is enabled by default. If a user initiates an invalid volume resize, for example by specifying a new size that is too big to be satisfied by the underlying storage system, expansion of PVC will continuously be retried and fail. With this new feature, such a PVC can now be edited to request a smaller size to unblock the PVC. The PVC can be monitored by watching .status.allocatedResourceStatuses and events on the PVC.

Job API for management by external controllers

• Support in the Job API for the managed-by mechanism graduated to beta and is enabled by default. This enables integration with external controllers like MultiKueue.

Improved scheduling performance

• The Kubernetes QueueingHint feature enhances scheduling throughput by preventing unnecessary scheduling retries. It’s achieved by allowing scheduler plugins to provide per-plugin callback functions that make efficient requeuing decisions.

Acknowledgements

As always, we want to thank all the Googlers that provide their time, passion, talent and leadership to keep making Kubernetes the best container orchestration platform. We would like to mention especially Googlers who helped drive the features mentioned in this blog [John Belamaric, Wojciech Tyczyński, Michelle Au, Matthew Cary, Aldo Culquicondor, Tim Hockin Maciej Skoczeń Michał Woźniak] and the Googlers who helped bring 1.32 to GKE in record time.

By Federico Bongiovanni, Benjamin Elder, and Sen Lu – Google Kubernetes Engine

Google Season of Docs announces results of 2024 program

Google Season of Docs is happy to announce the 2024 program results, including the project case studies.

Google Season of Docs is a grant-based program where open source organizations apply for US$5-15,000 to hire technical writers to complete documentation projects. At the end of the six-month documentation development phase, organizations submit a case study to outline the problems their documentation project was intended to solve, how they are measuring the success of their documentation project, and what they learned during the project. The case studies are publicly available and are intended to help other open source organizations learn best practices in open source documentation.

The 2024 Google Season of Docs documentation development phase began on April 10 and ended November 22, 2024. Participants in the 2024 program will also answer three followup surveys in 2025, in order to better track the impact of these documentation projects over time.

Key Takeaways from 2024 Google Season of Docs

Eleven different organizations participated in the 2024 program. These organizations represented a variety of open source project types, including databases, AI/ML, cloud infrastructure, programming languages, and science and medicine. The documentation projects hoped to address a variety of different problems. The most common challenges addressed by the Season of Docs projects were:

• Project documentation disorganized
• Potential project users having difficulty installing, using, or integrating your project
• Project documentation outdated
• Project documentation needs to be converted to a different tool, platform, or format
• Potential contributors having difficulty onboarding to your project
• Project documentation lacking for a specific key use case
• Potential project users lack fundamental understanding of the project domain

Program participants learned a lot from their projects. These lessons are detailed in the published case studies, to help other open source organizations who are interested in taking on their own documentation projects. Some highlights include:

“Putting time and effort into your project’s infrastructure, such as communication channels and onboarding processes, is really valuable work."

– CO2.js project

“Perhaps the key piece of advice that we came away with that could be useful for other projects is to be flexible in what you set out to accomplish: what looks like the top-five items on the to-do list on day one may not be what you think is the most important at the end."

– FontTools project

“This particular experiment with a different medium has turned out to be successful and we encourage other communities to also explore different media depending on their audience and information needs."

– Numpy project

“Developing documentation also helped us identify ambiguities in the interface and other areas of the site design or features that needed refinement or decisions made in order to document them properly. Now, with a base site and well-established documentation workflow, we are documenting more features as they are developed."

– R-Universe project

Take a look at the participant list to see the initial project plans and case studies for all of the participating projects!

What’s next?

Stay tuned for information about Google Season of Docs 2025 – watch for posts on this blog and sign up for the announcements email list.

By Elena Spitzer and Erin McKean, Google Open Source Programs Office

BazelCon 2024: A celebration of community and the launch of Bazel 8

The Bazel community celebrated a landmark year at BazelCon 2024. With a record-breaking 330+ attendees, 125+ talk proposal submissions, and a renewed focus on community-driven development, BazelCon marked a significant step forward for the build system and its users.

BazelCon 2024: Key highlights

The 8th annual build conference was held at the Computer History Museum in Mountain View, CA, on October 14 - 15, 2024. This was the first BazelCon not solely organized by Google; instead, it was organized by The Linux Foundation together with sponsors Google, BuildBuddy, EngFlow, NativeLink, AspectBuild, Gradle, Modus Create, and VirtusLab. The conference welcomed build enthusiasts from around the world to explore the latest advancements in build technologies, share learnings, and connect with each other.

The conference kicked off with an opening keynote delivered by Mícheál Ó Foghlú and Tobias Werth (Google), Alex Eagle (Aspect Build Systems), Helen Altshuler (EngFlow), and Chuck Grindel (Reveal Technology). The keynote highlighted the vital role of community contributions and charted a course for a future where Bazel thrives through shared stewardship.

Following the keynote, John Field and Tobias Werth (Engineering Managers at Google) delivered a state-of-the-union address, celebrating the year's top contributors and highlighting key achievements within the Bazel ecosystem.

Over the course of the conference, members of the Bazel community showcased their expertise and shared key insights through a series of live presentations. Some highlights include:

• Spotify's compelling Bazel adoption journey
• EngFlow's insightful post-mortems on remote execution
• Explorations of cutting-edge features like BuildBuddy's "Remote Bazel”

Take a look at our playlist of BazelCon 2024 Talks at your convenience.

In addition to main stage talks, BazelCon provided ample opportunities for attendees to connect and collaborate. Birds of a Feather sessions fostered lively discussions on topics ranging from generating SBOM using Bazel, to IDE integrations, to external dependency management, allowing community members to provide direct feedback and shape the future of Bazel. Make sure to check out raw BazelCon '24 Birds of a Feather notes from these sessions.

BazelCon 2024 also served as the launchpad for Bazel 8, a long-term support (LTS) release that brings significant enhancements to modularity, performance, and dependency management.

What’s new in Bazel 8?

• Starlark-powered modularity: Many core language rules traditionally shipped with Bazel are now Starlarkified and split into their own modules, including all Android, Java, Protocol Buffers, Python, and shell rules.
• WORKSPACE deprecation: The legacy WORKSPACE mechanism for external dependency management is disabled by default in Bazel 8, and is slated for removal in Bazel 9. Bzlmod, the default since Bazel 7, is the recommended solution going forward.
• Symbolic macros: Bazel 8 introduces a new way to write macros for build files, addressing many pitfalls and footguns of legacy macros. Symbolic macros offer better visibility encapsulation, type safety, and are amenable to lazy evaluation, coming in a future Bazel release.

Read the full release notes for Bazel 8.

Stay connected with the Bazel community

We extend our gratitude to everyone that contributed to the success of BazelCon 2024! We look forward to seeing you again next year.

To stay informed about the latest developments in the Bazel world, connect with us through the following channels:

• Bazel blog
• GitHub
• Slack
• Mailing list
• X (Twitter)

We encourage you to share your Bazel projects and experiences with us at product@bazel.build. We're always excited to hear from you!

By Keerthana Kumar and Xudong Yang, on behalf of the Google Bazel Team

A Robust Open Ecosystem for All: Accelerating AI Infrastructure

JAX now runs on AWS Trainium: Open Source Fuels AI Innovation

Open source software is the foundation of machine learning. It accelerates innovation through an ethos of flexibility and collaboration. This philosophy drives the open development of JAX, our high-performance array computing library, as well as OpenXLA, the compiler and runtime infrastructure it relies on.

Today we're excited to highlight how this commitment to openness, together with JAX and OpenXLA's modular designs, enables seamless integration of AWS Trainium and Trainium2 chips accelerators into the JAX ecosystem. Users get more portability, more choice, and faster progress.

JAX and OpenXLA, abstraction and modularity

JAX is a Python library for high-performance, large-scale numerical computing and machine learning. Its unique compiler-oriented design makes numerical computation familiar and portable while also accelerator-friendly and scalable. It combines a NumPy-like API with composable transformations for automatic differentiation, vectorization, parallelization, and more. Under the hood, JAX leverages the XLA compiler to optimize and scale computations over a broad set of backends.

This abstraction layer is key to its portability: JAX presents a consistent interface while XLA optimizes performance, whether you're running on CPUs, GPUs, TPUs, or something new.

In fact, OpenXLA infrastructure is designed to be modular and extensible to new platforms. By developing a PJRT plugin and leveraging existing XLA compiler components, JAX code can target new platforms, even when scaling from a single device to thousands.

Enter AWS Trainium and Inferentia

We are excited to announce that AWS Trainium is the latest platform to embrace JAX and OpenXLA. With the JAX Neuron plugin, AWS Trainium and Inferentia can be used as native JAX devices.

This new backend demonstrates how abstraction and modularity make JAX and OpenXLA especially extensible and amenable to collaboration, even on new hardware. We're thrilled to have diverse hardware partners like AMD, Arm, Intel, Nvidia, and AWS taking advantage of JAX's portability and performance. If you're interested in bringing new platforms into the JAX and OpenXLA ecosystem, please reach out!

A multi-platform ecosystem fosters open collaboration in advancing AI infrastructure. Our goal is to drive continuous development of open standards and to accelerate progress. And if you're a machine learning developer or numerical computing user, we're excited for you to try JAX on any platform you choose.

By Matthew Johnson - Principal Scientist, with additional contributors: Aditi Joshi, Fenghui Zhang, Roy Frostig, and Carlos Araya