v1.58.7

Added

• Allow incremental configurations for default job constraints, from @nsinkov

v1.58.6

Added

• Field for command length, from @dposada
• Logging of job instance when rebalancer preemption transaction fails, from @dposada

v1.58.5

Added

• Metrics for node and pod counts, from @nsinkov
• :production? to the config, from @dposada

Changed

• Changed logging from ERROR to INFO when a deleted cluster's watch fails, from @dposada

v1.58.4

Changed

• Changed logging with finalizer deletion, from @scrosby
• Collected metrics for waiting jobs under quota, from @nsinkov

v1.58.3

Changed

• Cleaned up the 'no acceptable compute cluster' log, from @dposada

v1.58.2

Changed

• Added support for transforming job constraints via configuration, from @dposada

Fixed

• Filtering out unsound GPU nodes, from @dposada

v1.58.1

Changed

• Support upserting > 1 incremental config in one transaction, from @nsinkov

Fixed

• Fix compile error on ex-info call, from @scrosby
• Fix docker environment to work with clojure 1.10, from @scrosby

v1.58.0

Changed

• Switch to JDK-11 and Clojure 1.10, from @scrosby

Added

• Cook can add a finalizer to pods, from @scrosby
• Support for incremental config for checkpointing volume mounts, from @nsinkov
• Support for rotating logs every hour, from @scrosby

v1.57.1

Changed

• Trimmed down pod event logging, from @dposada

v1.57.0

Added

• Optional comments to incremental value configurations, from @nsinkov

Changed

• Trimmed down pod metadata logging, from @dposada

Fixed

• Reverting JDK11 upgrade (back to JDK8), from @nsinkov