Skip to content

starting lots of workers/schedulers in tests starts a lot of _measure_tick PeriodicCallbacks #5908

@graingert

Description

@graingert

_measure_tick runs by default every 20ms for every distributed.core.Server created, eg every Worker and Scheduler

def _measure_tick(self):
now = time()
diff = now - self._last_tick
self._last_tick = now
if diff > tick_maximum_delay:
logger.info(
"Event loop was unresponsive in %s for %.2fs. "
"This is often caused by long-running GIL-holding "
"functions or moving large chunks of data. "
"This can cause timeouts and instability.",
type(self).__name__,
diff,
)
if self.digests is not None:
self.digests["tick-duration"].add(diff)

this can be a particular problem in tests such as test_adapt_quickly that starts 10 workers and a scheduler and so has 11 of these calls every 20ms

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions