AnchorRecomputation

Manual Anchor Date recomputation after a group configuration change.

📄️Resume a stalled recomputation run

Re-enqueues the next batch of a run whose chain broke, continuing from its cursor. Idempotent: the cursor decides where work restarts, so resuming twice — or resuming a run that turned out to be healthy — cannot skip or repeat members. A finished run is left alone.

📄️Recompute every member's Anchor Date

Queues a recomputation of the Anchor Date of every practitioner in the group, applying the group's current configuration to the members already in it. Returns as soon as the run is recorded — the work runs in sequential batches behind Cloud Tasks, and the returned runId is how it is followed. Safe to call twice: each call starts its own run, and a run writes nothing where a practitioner's Anchor Date has not moved.