0.17.0

This page documents the changes in Kannika Armory version 0.17.0.

Changelog 0.17.0

Features

• Operator: add a pod reaper that deletes managed backup pods stuck in ContainerCreating so their controller can reschedule them onto a healthy node. It is enabled by default and configurable through the operator’s Helm values. Reaping is skipped for permanent waiting reasons such as ImagePullBackOff and CreateContainerConfigError.

Console

• Restore: select multiple restores on the overview and delete them in bulk.

• Schema registry restore: select multiple schema registry restores on the overview and delete them in bulk.

• Restore: show a collapsible syntax help panel with an example and documentation link in the schema mapping and topics editors, plus a placeholder example in empty editors.

• Redirect to the relevant overview page when a resource being viewed no longer exists.

API

• Reworked database access threading: read queries run on a dedicated read pool with a single serialized writer, removing the contention that could make the overview slow or time out under load.

Helm Charts

• Add operator.config.podReaper.enabled value (default: true).

• Add operator.config.podReaper.timeout value (default: 5m), the time a pod may stay stuck in ContainerCreating before it is reaped.

• Add operator.config.podReaper.interval value (default: 1m), how often the operator scans for stuck pods.

• Grant the operator RBAC permission to emit Kubernetes Events for pod reaping.

Documentation

• Added Pod Reaper configuration documentation.

• Added FAQ entries on restored offsets, timestamps, and consumer groups.

• Added Upgrading to 0.17.x guide.

Miscellaneous Tasks

• Bump version to 0.17.0.