When the HBI outbox connector fails after a database change, it requires manual intervention to restart. We need a way for teams to restart connectors without needing direct cluster access. Options to explore include using the strimzi.io/restart annotation on the KafkaConnector resource and exposing that through app-interface, or adding a stop/start workflow via app-interface PRs. This came up during a stage incident on 2026-03-06 where the connector died after an RDS change and had to be restarted via the connect API.