hpcflow.sdk.monitoring.scheduler_cache.query_snapshot#

hpcflow.sdk.monitoring.scheduler_cache.query_snapshot(app, context, references, fetch, min_interval, ignore_interval=False, progress=None, scheduler_name='', no_update=False)#

Query at most once per interval, including across concurrent show processes.

Cache failures too, so repeated invocations cannot hammer an unavailable scheduler. A per-target OS lock covers the cache check and scheduler call. The override bypasses the age check, while zero disables the cache entirely. no_update reuses existing snapshots without writing new ones or lock files. A live query in that mode cannot refresh the shared rate-limit timestamp.

Parameters:
Return type:

SchedulerSnapshot