Added
- Added
TERMINALS_IDLE_CLEANUP_TIMEOUT_SECONDSso deployments can choose how long idle cleanup waits for one terminal stop or reset before trying again later.
Changed
- After a Terminals restart, recovered Docker and Kubernetes terminals now keep their policy settings instead of falling back to the global defaults.
- After a Terminals restart, recovered terminals keep their last recorded activity time instead of being treated as newly active.
- New Kubernetes terminal pods now save the real policy id, so recovered pods are matched to the right policy even when the policy label is shortened.
Fixed
- Fixed idle cleanup getting stuck behind one slow terminal stop or reset. Timed-out cleanup is left tracked and retried on the next sweep.