v0.17.0

There should be no breaking change. 🎉

• Add clusterSemver check plugin.
• Setting the cloud.sap/maintenance-log-details label to true on a node, enables detailed logging for that node.
• The ESX controller now exports the underlying ESX version in the cloud.sap/esx-version label.
• Refacotring.
• Update dependencies.

v0.16.3

There should be no breaking change. 🎉

• The state of the previous reconciliation of a node is tracked properly per profile. Beforehand it was tracked globally and only updated when notifying, which caused confusing behavior with NotifyPeriodic.
• Update dependencies.

v0.16.2

There should be no breaking change. 🎉

• Fix panic in TouchShuffles if resources are owned by unexpected resources

v0.16.1

There should be no breaking change. 🎉

• Shuffle metrics now get zero initalized
• Failures in shuffle metric collection no longer block state transitions

v0.16.0

Breaking changes 💥:

• Maintenance profiles are now executed concurrently. This is only constrained by allowing at max one profile to put a node into maintenance at any given time. The cloud.sap/maintenance-state label is now only informative. The actual relevant state data is stored within the cloud.sap/maintenance-data annotation. Also profile removals are now executed instantly instead of halting all profiles if the node was non-operational.

Other changes 🎉:

• Add a log message, when the controller really exits.
• Update dependencies.

v0.15.1

There should be no breaking change. 🎉

• Use semver for versioning
• Fix that config changes to the slackThread and stagger plugins did not update their respective lease durations, which in turn caused no effect for such changes.

v0.15

There should be no breaking change. 🎉

• The maintenance-controller is perfectly capable of draining itself. As shuffle metrics are generated, when moving a node into maintenance, the controller likely terminates before the updated metrics get scraped. Now on SIGTERM the controller waits for an other scrape.

v0.14

There should be no breaking change. 🎉

• Fix quadratic notification bug
• Use RFC3339 for log timestamps

v0.13

There should be no breaking change. 🎉

• Added pod shuffle prometheus metrics.
• Added waitExclude check plugin
• Updated Dockerimage to golang 1.18

v0.12

Breaking changes 💥:

• Changed the format of ./config/maintenance.yaml. Check the readme for further information.
• Made state transitions configurable.
• Added notification schedules. The old implicit behavior is reflected by the periodic schedule.

Other changes 🎉:

• Updated dependencies