test(scheduler): runs started in a row start in that order

The scheduler tests open several runs within microseconds and read them back newest first, by start time. A container's clock on a VM can step back between two readings, which sorted the third of three runs before the second; the tests now run on a clock that never repeats or goes back.

backend/tests/test_scheduler.py

@@ -22,6 +22,31 @@ from app.runstatus import RunStatus, RunTrigger
 _JOB = "unit_test_claim"  # not a real job id — exercises the primitive without touching live jobs
 
 
+class _StrictlyLater(datetime):
+    """A clock that never repeats or goes back: each reading at least a microsecond after the last."""
+
+    last: datetime | None = None
+    clock: type[datetime] = datetime
+
+    @classmethod
+    def now(cls, tz=None):
+        t = cls.clock.now(tz)
+        if cls.last is not None and t <= cls.last:
+            t = cls.last + timedelta(microseconds=1)
+        cls.last = t
+        return t
+
+
+@pytest.fixture(autouse=True)
+def _runs_start_in_order(monkeypatch):
+    """Runs started in a row start in that order. These tests open several runs within microseconds
+    and read them back newest first, by start time; the wall clock of a container on a VM can step
+    back between two readings (seen: the third of three runs sorted before the second)."""
+    _StrictlyLater.last = None
+    _StrictlyLater.clock = jobruns.datetime
+    monkeypatch.setattr(jobruns, "datetime", _StrictlyLater)
+
+
 def _release():
     scheduler._record(_JOB, running=False)