Provide better return indicating stale scheduler

[gisjedi]

This should not return a 204. Instead it should return something more indicative of the scheduler being stale.

Version and Environment Details

• Scale API: 7.0.1rc4
  • Can be found at /api/v6/scheduler/status
• Scale UI: 0.8.4
  Cluster details below are very helpful for problem resolution, but external Scale users aren't expected to provide.
• DC/OS: kraken

Additional context

scale/scale/scheduler/views.py

Lines 145 to 146 in 330b6e6

	if (now() - status_timestamp).total_seconds() > StatusView.STATUS_FRESHNESS_THRESHOLD:
	return Response(status=status.HTTP_204_NO_CONTENT)

[JohnPTobe]

Return the status but with a field warning that it's stale? Return only the stale warning with no status? I assume the problem was you tore your hair out debugging the scheduler only to figure out it was slow so something more obvious would be nice.

[cshamis]

That bug was also complicated by some sort of hang up in the node deprecate code that confused the scheduler.

…

On Sep 6, 2019, at 5:51 PM, John Tobe ***@***.***> wrote: Return the status but with a field warning that it's stale? Return only the stale warning with no status? I assume the problem was you tore your hair out debugging the scheduler only to figure out it was slow so something more obvious would be nice. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.