The operator console now surfaces one health signal per subsystem, and every scheduled job declares its own cadence so minute-long tasks run in the worker queue instead of the request cycle.
What changed
- Health signals for webhooks, queues, tokens, quality ratings and storage in one place.
- Cron steps with heavy work moved to the background worker.
- Alerts throttle themselves so an incident produces one notification, not fifty.
What to do
Nothing is required. The change applies to every workspace. If you notice a behaviour that differs from a help article, the article date tells you whether it predates this release; write to [email protected] and we will correct it.
