=== PART 4 — double failure on the SHIPPED 0.221.1
TRIGGER (deliberate, realistic): the undo copy is LOST between being written and being needed —
  a drive that goes away, a filesystem that goes read-only, or an external cleanup. The product
  must not assume the file it wrote twenty seconds ago is still there.
  TRIGGER: removed the freshly-written undo copy pre-restore-20260822T221623Z-docmost-postgres.sql.tmp
=== app state — must be NOT running:
docmost | Up 12 minutes (healthy)
docmost-postgres | Up 12 minutes (healthy)
docmost-redis | Up 12 minutes (healthy)
