The community around Kubernetes includes a number of Special Interest Groups (SIGs) and Working Groups (WGs) facilitating discussions on important topics between interested contributors. Today we would like to announce the new Kubernetes Checkpoint Restore WG focusing on the integration of Checkpoint/Restore functionality into Kubernetes.
Motivation and use cases There are several high-level scenarios discussed in the working group:
Optimizing resource utilization for interactive workloads, such as Jupyter notebooks and AI chatbots Accelerating startup of applications with long initialization times, including Java applications and LLM inference services Using periodic checkpointing to enable fault-tolerance for long-running workloads, such as distributed model training Providing interruption-aware scheduling with transparent checkpoint/restore, allowing lower-priority Pods to be preempted while preserving the runtime state of applications Facilitating Pod migration across nodes for load balancing and maintenance, without disrupting workloads.

K8S Contributors

Kubernetes announces a new Checkpoint/Restore Working Group focused on integrating checkpoint/restore functionality into Kubernetes. The group addresses use cases including optimizing resource utilization for interactive workloads, accelerating startup times for applications with long initialization, enabling fault-tolerance through periodic checkpointing, supporting interruption-aware scheduling, facilitating pod migration, and enabling forensic analysis. The working group bridges the Kubernetes community with the CRIU ecosystem, which includes tools like CRIU, checkpointctl, criu-coordinator, and checkpoint-restore-operator. Contributors can participate through bi-weekly meetings, Slack channels, and mailing lists.

Announcing the Checkpoint/Restore Working Group