Remote connections now reconnect in a controlled way

A short network interruption should not pull BloomMD out of the user's workflow. The desktop client and SyncProvider now share one clearly defined reconnect flow.

One flow instead of competing loops

Temporary interruptions use exponential backoff with a bounded number of attempts. Each new attempt can request a fresh sync ticket, so expired tickets are not reused indefinitely.

Permanent states such as revoked access, a missing document or a full room switch to a clear failure state. From there, the connection can be restarted manually.

The interface distinguishes offline, reconnecting and a permanently failed reconnect. Intentional disconnects and workspace changes do not trigger a late reconnect.

Commit date: August 28, 2026 · View the release