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 analogous 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 transition(self, key, start, finish, *args, **kwargs):
...         if start == 'processing' and finish == 'memory':
...             self.counter += 1
...
...     def restart(self, scheduler):
...         self.counter = 0

>>> plugin = Counter()
>>> scheduler.add_plugin(plugin)  # doctest: +SKIP

add_client(scheduler=None, client=None, **kwargs)

Run when a new client connects

add_worker(scheduler=None, worker=None, **kwargs)

Run when a new worker enters the cluster

close()

Run when the scheduler closes down

This runs at the beginning of the Scheduler shutdown process, but after workers have been asked to shut down gracefully

remove_client(scheduler=None, client=None, **kwargs)

Run when a client disconnects

remove_worker(scheduler=None, worker=None, **kwargs)

Run when a worker leaves the cluster

restart(scheduler, **kwargs)

Run when the scheduler restarts itself

start(scheduler)

Run when the scheduler starts up

This runs at the end of the Scheduler startup process

transition(key, start, finish, *args, **kwargs)

Run whenever a task changes state

Parameters:

key: string

start: string

Start state of the transition. One of released, waiting, processing, memory, error.

finish: string

Final state of the transition.

*args, **kwargs: More options passed when transitioning

This may include worker ID, compute time, etc.

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

Run when a new graph / tasks enter the scheduler

RabbitMQ Example

RabbitMQ is a distributed messaging queue that we can use to post updates about task transitions. By posting transitions to RabbitMQ, we allow other machines to do the processing of transitions and keep scheduler processing to a minimum. See the RabbitMQ tutorial for more information on RabbitMQ and how to consume the messages.

import json
from distributed.diagnostics.plugin import SchedulerPlugin
import pika

class RabbitMQPlugin(SchedulerPlugin):
    def __init__(self):
        # Update host to be your RabbitMQ host
        self.connection = pika.BlockingConnection(
            pika.ConnectionParameters(host='localhost'))
        self.channel = self.connection.channel()
        self.channel.queue_declare(queue='dask_task_status', durable=True)

    def transition(self, key, start, finish, *args, **kwargs):
        message = dict(
            key=key,
            start=start,
            finish=finish,
        )
        self.channel.basic_publish(
            exchange='',
            routing_key='dask_task_status',
            body=json.dumps(message),
            properties=pika.BasicProperties(
                delivery_mode=2,  # make message persistent
            ))

@click.command()
def dask_setup(scheduler):
    plugin = RabbitMQPlugin()
    scheduler.add_plugin(plugin)

Run with: dask-scheduler --preload <filename.py>

Accessing Full Task State

If you would like to access the full distributed.scheduler.TaskState stored in the scheduler you can do this by passing and storing a reference to the scheduler as so:

from distributed.diagnostics.plugin import SchedulerPlugin

class MyPlugin(SchedulerPlugin):
    def __init__(self, scheduler):
         self.scheduler = scheduler

    def transition(self, key, start, finish, *args, **kwargs):
         # Get full TaskState
         ts = self.scheduler.tasks[key]

@click.command()
def dask_setup(scheduler):
    plugin = MyPlugin(scheduler)
    scheduler.add_plugin(plugin)

Worker Plugins

distributed.diagnostics.plugin.WorkerPlugin provides a base class for creating your own worker plugins. In addition, Dask provides some built-in plugins.

class distributed.diagnostics.plugin.WorkerPlugin

Interface to extend the Worker

A worker plugin enables custom code to run at different stages of the Workers’ lifecycle: at setup, during task state transitions, when a task or dependency is released, and at teardown.

A plugin enables custom code to run at each of step of a Workers’s life. Whenever such an event happens, the corresponding method on this class will be called. Note that the user code always runs within the Worker’s main thread.

To implement a plugin implement some of the methods of this class and register the plugin to your client in order to have it attached to every existing and future workers with Client.register_worker_plugin.

Examples

>>> class ErrorLogger(WorkerPlugin):
...     def __init__(self, logger):
...         self.logger = logger
...
...     def setup(self, worker):
...         self.worker = worker
...
...     def transition(self, key, start, finish, *args, **kwargs):
...         if finish == 'error':
...             exc = self.worker.exceptions[key]
...             self.logger.error("Task '%s' has failed with exception: %s" % (key, str(exc)))

>>> plugin = ErrorLogger()
>>> client.register_worker_plugin(plugin)  # doctest: +SKIP

release_dep(dep, state, report)

Called when the worker releases a dependency.

Parameters:

dep: string

state: string

State of the released dependency. One of waiting, flight, memory.

report: bool

Whether the worker should report the released dependency to the scheduler.

release_key(key, state, cause, reason, report)

Called when the worker releases a task.

Parameters:

key: string

state: string

State of the released task. One of waiting, ready, executing, long-running, memory, error.

cause: string or None

Additional information on what triggered the release of the task.

reason: None

Not used.

report: bool

Whether the worker should report the released task to the scheduler.

setup(worker)

Run when the plugin is attached to a worker. This happens when the plugin is registered and attached to existing workers, or when a worker is created after the plugin has been registered.

teardown(worker)

Run when the worker to which the plugin is attached to is closed

transition(key, start, finish, **kwargs)

Throughout the lifecycle of a task (see Worker), Workers are instructed by the scheduler to compute certain tasks, resulting in transitions in the state of each task. The Worker owning the task is then notified of this state transition.

Whenever a task changes its state, this method will be called.

Parameters:

key: string

start: string

Start state of the transition. One of waiting, ready, executing, long-running, memory, error.

finish: string

Final state of the transition.

kwargs: More options passed when transitioning

Built-In Worker Plugins

class distributed.diagnostics.plugin.PipInstall(packages, pip_options=None, restart=False)

A Worker Plugin to pip install a set of packages

This accepts a set of packages to install on all workers. You can also optionally ask for the worker to restart itself after performing this installation.

Note

This will increase the time it takes to start up each worker. If possible, we recommend including the libraries in the worker environment or image. This is primarily intended for experimentation and debugging.

Additional issues may arise if multiple workers share the same file system. Each worker might try to install the packages simultaneously.

Parameters:

packages : List[str]

A list of strings to place after “pip install” command

pip_options : List[str]

Additional options to pass to pip.

restart : bool, default False

Whether or not to restart the worker after pip installing Only functions if the worker has an attached nanny process

Examples

>>> from dask.distributed import PipInstall
>>> plugin = PipInstall(packages=["scikit-learn"], pip_options=["--upgrade"])

>>> client.register_worker_plugin(plugin)