- Fixed an issue where connections on non-English instances could be incorrectly marked Down by the health monitor even though the connection was working normally. (PRB2050565)

- Fixed an issue where the connection health scan check incorrectly reported a connection issue when the instance language was set to something other than English. (PRB2056925)

- Fixed a security issue where an authenticated user could access internal integration operations without the proper authorization, including retrieving OAuth credentials. (PRB2028445)

- Fixed a security issue where the Health Dashboard's scan check panel could be accessed without proper authorization. (PRB2038484)

- Reworked connection health monitoring (Heartbeat) to reduce false "Slow" and "Down" alerts on busy connections.

- Improved the Restart Connection experience — a clear banner now appears on the connection form when a connection is down and repairable, with confirmation messaging on success or a clear error if you lack access.

- Reduced duplicate Issue records on the Health Dashboard — recurring errors with the same cause, source, and connection now update the existing Issue instead of creating a new one, and unknown errors now get more descriptive short descriptions.

- Improved recovery for slow (not down) connections — inbound message processing now gets progressively more time before timing out, instead of failing repeatedly at a fixed 30-second limit.

- Added a Category field to Known Error Codes, now visible on the Health Dashboard issue list and on Issue records and lists, to help you quickly identify the type of problem (onboarding, data, configuration, etc.) behind a health finding.