sync-v2 step 7: observability + graceful shutdown (§7)

- Heartbeat JSON now carries {owned, sockets_open, tiers:{hot,warm,cold}, ts} per
  §2 so peers and operators can see each worker's load and socket footprint.
- run.py runs the sync coordinator directly (not via a placeholder arq worker) so
  SIGINT/SIGTERM can stop the claimer, cancel runners, and release every lease
  explicitly within a 10s drain budget — peers adopt the accounts immediately
  instead of waiting out the lease TTL. Removed the sync:control-queue hack and
  the supervisor indirection (start/stop_supervisor, sync_on_startup/shutdown).
- Acceptance: graceful-shutdown test proves a peer adopts released accounts within
  seconds under a 30s TTL (i.e. via release, not expiry).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

backend/tests/sync/test_acceptance.py

@@ -297,6 +297,35 @@ async def test_event_driven_teardown(connections, make_accounts) -> None:  # typ
         await c.stop()
 
 
+# --- acceptance 6: graceful shutdown -----------------------------------------
+
+
+async def test_acceptance_6_graceful_shutdown(connections, make_accounts) -> None:  # type: ignore[no-untyped-def]
+    """A graceful stop releases every lease so a peer adopts the accounts fast.
+
+    Leases are long-lived, so adoption within a couple of seconds proves the
+    accounts were handed over by explicit release — not by waiting out the TTL.
+    """
+    ids = set(await make_accounts(6))
+    await _seed(ids)  # hot => persistent IDLE
+    c1 = _fast(SyncCoordinator("w1"))
+    c2 = _fast(SyncCoordinator("w2"))
+    c1.lease_ttl = c2.lease_ttl = 30  # adoption must come from release, not expiry
+    await c1.start()
+    await _until(lambda: c1.owned == ids)  # sole owner
+    await c2.start()  # nothing to claim yet — all leased by w1
+    try:
+        await c1.stop()  # graceful: stops claimer, cancels runners, releases leases
+        assert c1.owned == set()
+        assert await get_redis().exists(leases.worker_key("w1")) == 0  # heartbeat gone
+        # w2 adopts every account well within the 30s TTL, one connection each.
+        await _until(lambda: c2.owned == ids, timeout_s=4.0)
+        await _until(lambda: all(connections[a] == 1 for a in ids))
+    finally:
+        await c2.stop()
+    await _until(lambda: all(connections[a] == 0 for a in ids))
+
+
 # --- §5 activity tiers -------------------------------------------------------