Scheduler Plugins

class distributed.diagnostics.plugin.SchedulerPlugin

Interface to extend the Scheduler

The scheduler operates by triggering and responding to events like task_finished, update_graph, task_erred, etc..

A plugin enables custom code to run at each of those same events. The scheduler will run the analagous methods on this class when each event is triggered. This runs user code within the scheduler thread that can perform arbitrary operations in synchrony with the scheduler itself.

Plugins are often used for diagnostics and measurement, but have full access to the scheduler and could in principle affect core scheduling.

To implement a plugin implement some of the methods of this class and add the plugin to the scheduler with Scheduler.add_plugin(myplugin).

Examples

>>> class Counter(SchedulerPlugin):
...     def __init__(self):
...         self.counter = 0
...
...     def task_finished(self, scheduler, key, worker, nbytes):
...         self.counter += 1
...
...     def restart(self, scheduler):
...         self.counter = 0

>>> c = Counter()
>>> scheduler.add_plugin(c)  

restart(scheduler, **kwargs)

Run when the scheduler restarts itself

task_erred(scheduler, key=None, worker=None, exception=None, **kwargs)

Run when a task is reported failed

task_finished(scheduler, key=None, worker=None, nbytes=None, **kwargs)

Run when a task is reported complete

update_graph(scheduler, dsk=None, keys=None, restrictions=None, **kwargs)

Run when a new graph / tasks enter the scheduler