v1.51.1

Changed

• Support ignoring specific group ID's when computing supplemental group IDs
• Log exceptions in deep-merge-with

Fixed

• Take only the top X pending jobs when triggering k8s autoscaling

v1.51.0

Added

• Support for longer pod names
• Configurable validation of job resources by node type
• /usage for all users

Changed

• Make job->acceptable-compute-clusters configurable

v1.50.1

Added

• Ability to not set memory limits
• Authenticator refresh logic needed for non-GKE k8s
• Checkpoint locality constraint

Changed

• Logging the largest job and offer by resource

v1.50.0

• Adds location to compute cluster
• Schedules and matches jobs with disk
• Makes k8s API client read timeout configurable

v1.49.2

• Add resource request and limit to init-container in pod
• Refactor authentication initialization
• Migrate to GitHub Actions from Travis CI
• Allow users to use int values for disk request and disk limit

v1.49.1

update changelog and version for v1.49.1

v1.49.0

update changelog and version for v1.49.0

v1.48.4

Changed

• Reduced excessive logging for checkpointing and launching tasks

v1.48.3

Changed

• Added supplemental groups to the pod security context

v1.48.2

Changed

• Reduced excessive logging for k8s dynamic clusters and writing tasks