fix(mcp): get_anomaly shows a quick Steward restart, not only a gap over 90 s (#5862)

A redeploy's new process ticks at once, so its once-a-minute samples never
broke for 90 s and get_anomaly showed none of the day's restarts. The first
self-monitor tick in a process now records restart_gap_secs (time since the
last process's last sample), and steward_gaps shows each one as a restart.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

tests/core/test_self_monitor.py

@@ -144,3 +144,25 @@ def test_unwatched_counts_only_a_real_gap():
     assert sm.unwatched_secs(None, now) == 0.0                          # first ever tick
     assert sm.unwatched_secs(now - timedelta(seconds=61), now) == 0.0   # an ordinary minute
     assert sm.unwatched_secs(now - timedelta(minutes=7), now) == 420.0  # a deploy or outage
+
+
+def test_a_fresh_process_marks_its_restart_once(monkeypatch):
+    """#5862: the first tick records how long since the last process's last
+    sample, however short; later ticks and a first-ever start don't."""
+    from datetime import datetime, timedelta, timezone
+    recorded = _patch_recorder(monkeypatch)
+    monkeypatch.setattr(sm, "_started", False)
+    monkeypatch.setattr(_FakeSession, "last", datetime.now(timezone.utc) - timedelta(seconds=40))
+
+    asyncio.run(sm.record_self_metrics(_FakeApp()))
+    asyncio.run(sm.record_self_metrics(_FakeApp()))
+
+    gaps = [v for (_s, _r, m, v) in recorded if m == "restart_gap_secs"]
+    assert len(gaps) == 1 and 39 <= gaps[0] <= 60
+    assert dict((m, v) for (_s, _r, m, v) in recorded)["unwatched_secs"] == 0.0
+
+    recorded.clear()
+    monkeypatch.setattr(sm, "_started", False)
+    monkeypatch.setattr(_FakeSession, "last", None)
+    asyncio.run(sm.record_self_metrics(_FakeApp()))
+    assert "restart_gap_secs" not in {m for (_s, _r, m, _v) in recorded}