API reference

Data structures

class apscheduler.Task(*, id, func, executor, max_running_jobs=None, misfire_grace_time=None, state=None)

Represents a callable and its surrounding configuration parameters.

Variables:
  • id (str) – the unique identifier of this task

  • func (Callable) – the callable that is called when this task is run

  • max_running_jobs (int | None) – maximum number of instances of this task that are allowed to run concurrently

  • misfire_grace_time (timedelta | None) – maximum number of seconds the run time of jobs created for this task are allowed to be late, compared to the scheduled run time

class apscheduler.Schedule(*, id, task_id, trigger, args=(), kwargs=(), coalesce=CoalescePolicy.latest, misfire_grace_time=None, max_jitter=None, tags=(), next_fire_time=None, last_fire_time=None, acquired_by=None, acquired_until=None)

Represents a schedule on which a task will be run.

Variables:
  • id (str) – the unique identifier of this schedule

  • task_id (str) – unique identifier of the task to be run on this schedule

  • args (tuple) – positional arguments to pass to the task callable

  • kwargs (dict[str, Any]) – keyword arguments to pass to the task callable

  • coalesce (CoalescePolicy) – determines what to do when processing the schedule if multiple fire times have become due for this schedule since the last processing

  • misfire_grace_time (timedelta | None) – maximum number of seconds the scheduled job’s actual run time is allowed to be late, compared to the scheduled run time

  • max_jitter (timedelta | None) – maximum number of seconds to randomly add to the scheduled time for each job created from this schedule

  • tags (frozenset[str]) – strings that can be used to categorize and filter the schedule and its derivative jobs

  • conflict_policy (ConflictPolicy) – determines what to do if a schedule with the same ID already exists in the data store

  • next_fire_time (datetime) – the next time the task will be run

  • last_fire_time (datetime | None) – the last time the task was scheduled to run

  • acquired_by (str | None) – ID of the scheduler that has acquired this schedule for processing

  • acquired_until (str | None) – the time after which other schedulers are free to acquire the schedule for processing even if it is still marked as acquired

class apscheduler.Job(*, id=NOTHING, task_id, args=(), kwargs=(), schedule_id=None, scheduled_fire_time=None, jitter=NOTHING, start_deadline=None, result_expiration_time=datetime.timedelta(0), tags=(), created_at=NOTHING, started_at=None, acquired_by=None, acquired_until=None)

Represents a queued request to run a task.

Variables:
  • id (UUID) – autogenerated unique identifier of the job

  • task_id (str) – unique identifier of the task to be run

  • args (tuple) – positional arguments to pass to the task callable

  • kwargs (dict[str, Any]) – keyword arguments to pass to the task callable

  • schedule_id (str) – unique identifier of the associated schedule (if the job was derived from a schedule)

  • scheduled_fire_time (datetime | None) – the time the job was scheduled to run at (if the job was derived from a schedule; includes jitter)

  • jitter (timedelta | None) – the time that was randomly added to the calculated scheduled run time (if the job was derived from a schedule)

  • start_deadline (datetime | None) – if the job is started in the worker after this time, it is considered to be misfired and will be aborted

  • result_expiration_time (timedelta) – minimum amount of time to keep the result available for fetching in the data store

  • tags (frozenset[str]) – strings that can be used to categorize and filter the job

  • created_at (datetime) – the time at which the job was created

  • started_at (datetime | None) – the time at which the execution of the job was started

  • acquired_by (str | None) – the unique identifier of the worker that has acquired the job for execution

  • acquired_until (str | None) – the time after which other workers are free to acquire the job for processing even if it is still marked as acquired

property original_scheduled_time: datetime | None

The scheduled time without any jitter included.

Return type:

datetime | None

class apscheduler.JobInfo(*, job_id, task_id, schedule_id, scheduled_fire_time, jitter, start_deadline, tags)

Contains information about the currently running job.

This information is available in the thread or task where a job is currently being run, available from current_job.

Variables:
  • job_id (UUID) – the unique identifier of the job

  • task_id (str) – the unique identifier of the task that is being run

  • schedule_id (str | None) – the unique identifier of the schedule that the job was derived from (if any)

  • scheduled_fire_time (datetime | None) – the time the job was scheduled to run at (if the job was derived from a schedule; includes jitter)

  • jitter (timedelta) – the time that was randomly added to the calculated scheduled run time (if the job was derived from a schedule)

  • start_deadline (datetime | None) – if the job is started in the worker after this time, it is considered to be misfired and will be aborted

  • tags (frozenset[str]) – strings that can be used to categorize and filter the job

class apscheduler.JobResult(*, job_id, outcome, finished_at=NOTHING, expires_at, exception=None, return_value=None)

Represents the result of running a job.

Variables:
  • job_id (UUID) – the unique identifier of the job

  • outcome (JobOutcome) – indicates how the job ended

  • finished_at (datetime) – the time when the job ended

  • exception (BaseException | None) – the exception object if the job ended due to an exception being raised

  • return_value – the return value from the task function (if the job ran to completion successfully)

class apscheduler.RetrySettings(*, stop=<tenacity.stop.stop_after_delay object>, wait=<tenacity.wait.wait_exponential object>)

Settings for retrying an operation with Tenacity.

Parameters:
  • stop (stop_base) – defines when to stop trying

  • wait (wait_base) – defines how long to wait between attempts

Schedulers

class apscheduler.schedulers.sync.Scheduler(data_store=None, event_broker=None, *, identity=None, role=SchedulerRole.both, job_executors=None, default_job_executor=None, logger=None)

A synchronous scheduler implementation.

start_in_background()

Launch the scheduler in a new thread.

This method registers atexit hooks to shut down the scheduler and wait for the thread to finish.

Raises:

RuntimeError – if the scheduler is not in the stopped state

Return type:

None

property state: RunState

The current running state of the scheduler.

Return type:

RunState

subscribe(callback, event_types=None, *, one_shot=False)

Subscribe to events.

To unsubscribe, call the Subscription.unsubscribe() method on the returned object.

Parameters:
  • callback – callable to be called with the event object when an event is published

  • event_types – an iterable of concrete Event classes to subscribe to

  • one_shot – if True, automatically unsubscribe after the first matching event

class apscheduler.schedulers.async_.AsyncScheduler(*, data_store=NOTHING, event_broker=NOTHING, identity=None, role=SchedulerRole.both, max_concurrent_jobs=100, job_executors=None, default_job_executor=None, logger=<Logger apscheduler.schedulers.async_ (WARNING)>)

An asynchronous (AnyIO based) scheduler implementation.

Parameters:
  • data_store – the data store for tasks, schedules and jobs

  • event_broker – the event broker to use for publishing an subscribing events

  • max_concurrent_jobs – Maximum number of jobs the worker will run at once

  • role – specifies what the scheduler should be doing when running

  • process_schedulesTrue to process due schedules in this scheduler

async add_job(func_or_task_id, *, args=None, kwargs=None, job_executor=None, tags=None, result_expiration_time=0)

Add a job to the data store.

Parameters:
  • func_or_task_id (str | Callable) –

  • job_executor (str | None) – name of the job executor to run the task with

  • args (Iterable | None) – positional arguments to call the target callable with

  • kwargs (Mapping[str, Any] | None) – keyword arguments to call the target callable with

  • job_executor – name of the job executor to run the task with

  • tags (Iterable[str] | None) – strings that can be used to categorize and filter the job

  • result_expiration_time (timedelta | float) – the minimum time (as seconds, or timedelta) to keep the result of the job available for fetching (the result won’t be saved at all if that time is 0)

Return type:

UUID

Returns:

the ID of the newly created job

async add_schedule(func_or_task_id, trigger, *, id=None, args=None, kwargs=None, job_executor=None, coalesce=CoalescePolicy.latest, misfire_grace_time=None, max_jitter=None, tags=None, conflict_policy=ConflictPolicy.do_nothing)

Schedule a task to be run one or more times in the future.

Parameters:
  • func_or_task_id (str | Callable) – either a callable or an ID of an existing task definition

  • trigger (Trigger) – determines the times when the task should be run

  • id (str | None) – an explicit identifier for the schedule (if omitted, a random, UUID based ID will be assigned)

  • args (Iterable | None) – positional arguments to be passed to the task function

  • kwargs (Mapping[str, Any] | None) – keyword arguments to be passed to the task function

  • job_executor (str | None) – name of the job executor to run the task with

  • coalesce (CoalescePolicy) – determines what to do when processing the schedule if multiple fire times have become due for this schedule since the last processing

  • misfire_grace_time (float | timedelta | None) – maximum number of seconds the scheduled job’s actual run time is allowed to be late, compared to the scheduled run time

  • max_jitter (float | timedelta | None) – maximum number of seconds to randomly add to the scheduled time for each job created from this schedule

  • tags (Iterable[str] | None) – strings that can be used to categorize and filter the schedule and its derivative jobs

  • conflict_policy (ConflictPolicy) – determines what to do if a schedule with the same ID already exists in the data store

Return type:

str

Returns:

the ID of the newly added schedule

async get_job_result(job_id, *, wait=True)

Retrieve the result of a job.

Parameters:
  • job_id (UUID) – the ID of the job

  • wait (bool) – if True, wait until the job has ended (one way or another), False to raise an exception if the result is not yet available

Raises:

JobLookupError – if wait=False and the job result does not exist in the data store

Return type:

JobResult

async get_next_event(event_types)

Wait until the next event matching one of the given types arrives.

Parameters:

event_types – an event class or an iterable event classes to subscribe to

async get_schedule(id)

Retrieve a schedule from the data store.

Parameters:

id (str) – the unique identifier of the schedule

Raises:

ScheduleLookupError – if the schedule could not be found

Return type:

Schedule

async get_schedules()

Retrieve all schedules from the data store.

Returns:

a list of schedules, in an unspecified order

async remove_schedule(id)

Remove the given schedule from the data store.

Parameters:

id (str) – the unique identifier of the schedule

Return type:

None

async run_job(func_or_task_id, *, args=None, kwargs=None, job_executor=None, tags=())

Convenience method to add a job and then return its result.

If the job raised an exception, that exception will be reraised here.

Parameters:
  • func_or_task_id (str | Callable) – either a callable or an ID of an existing task definition

  • args (Iterable | None) – positional arguments to be passed to the task function

  • kwargs (Mapping[str, Any] | None) – keyword arguments to be passed to the task function

  • job_executor (str | None) – name of the job executor to run the task with

  • tags (Iterable[str] | None) – strings that can be used to categorize and filter the job

Return type:

Any

Returns:

the return value of the task function

async run_until_stopped(*, task_status=<anyio._core._tasks._IgnoredTaskStatus object>)

Run the scheduler until explicitly stopped.

Return type:

None

property state: RunState

The current running state of the scheduler.

Return type:

RunState

async stop()

Signal the scheduler that it should stop processing schedules.

This method does not wait for the scheduler to actually stop. For that, see wait_until_stopped().

Return type:

None

subscribe(callback, event_types=None, *, one_shot=False, is_async=True)

Subscribe to events.

To unsubscribe, call the Subscription.unsubscribe() method on the returned object.

Parameters:
  • callback – callable to be called with the event object when an event is published

  • event_types – an event class or an iterable event classes to subscribe to

  • one_shot – if True, automatically unsubscribe after the first matching event

  • is_asyncTrue if the (synchronous) callback should be called on the event loop thread, False if it should be called in a worker thread. If callback is a coroutine function, this flag is ignored.

async wait_until_stopped()

Wait until the scheduler is in the “stopped” or “stopping” state.

If the scheduler is already stopped or in the process of stopping, this method returns immediately. Otherwise, it waits until the scheduler posts the SchedulerStopped event.

Return type:

None

Workers

Data stores

class apscheduler.abc.DataStore

Asynchronous version of DataStore. Expected to work on asyncio.

abstract async acquire_jobs(worker_id, limit=None)

Acquire unclaimed jobs for execution.

This method claims up to the requested number of jobs for the given worker and returns them.

Parameters:
  • worker_id (str) – unique identifier of the worker

  • limit (int | None) – maximum number of jobs to claim and return

Return type:

list[Job]

Returns:

the list of claimed jobs

abstract async acquire_schedules(scheduler_id, limit)

Acquire unclaimed due schedules for processing.

This method claims up to the requested number of schedules for the given scheduler and returns them.

Parameters:
  • scheduler_id – unique identifier of the scheduler

  • limit – maximum number of schedules to claim

Returns:

the list of claimed schedules

abstract async add_job(job)

Add a job to be executed by an eligible worker.

Parameters:

job (Job) – the job object

Return type:

None

abstract async add_schedule(schedule, conflict_policy)

Add or update the given schedule in the data store.

Parameters:
  • schedule (Schedule) – schedule to be added

  • conflict_policy (ConflictPolicy) – policy that determines what to do if there is an existing schedule with the same ID

Return type:

None

abstract async add_task(task)

Add the given task to the store.

If a task with the same ID already exists, it replaces the old one but does NOT affect task accounting (# of running jobs).

Parameters:

task (Task) – the task to be added

Return type:

None

abstract async get_job_result(job_id)

Retrieve the result of a job.

The result is removed from the store after retrieval.

Parameters:

job_id (UUID) – the identifier of the job

Return type:

JobResult | None

Returns:

the result, or None if the result was not found

abstract async get_jobs(ids=None)

Get the list of pending jobs.

Parameters:

ids (Iterable[UUID] | None) – a specific set of job IDs to return, or None to return all jobs

Return type:

list[Job]

Returns:

the list of matching pending jobs, in the order they will be given to workers

abstract async get_next_schedule_run_time()

Return the earliest upcoming run time of all the schedules in the store, or None if there are no active schedules.

Return type:

datetime | None

abstract async get_schedules(ids=None)

Get schedules from the data store.

Parameters:

ids – a specific set of schedule IDs to return, or None to return all schedules

Returns:

the list of matching schedules, in unspecified order

abstract async get_task(task_id)

Get an existing task definition.

Parameters:

task_id (str) – ID of the task to be returned

Return type:

Task

Returns:

the matching task

Raises:

TaskLookupError – if no matching task was found

abstract async get_tasks()

Get all the tasks in this store.

Returns:

a list of tasks, sorted by ID

abstract async release_job(worker_id, task_id, result)

Release the claim on the given job and record the result.

Parameters:
  • worker_id (str) – unique identifier of the worker

  • task_id (str) – the job’s task ID

  • result (JobResult) – the result of the job

Return type:

None

abstract async release_schedules(scheduler_id, schedules)

Release the claims on the given schedules and update them on the store.

Parameters:
  • scheduler_id – unique identifier of the scheduler

  • schedules – the previously claimed schedules

abstract async remove_schedules(ids)

Remove schedules from the data store.

Parameters:

ids (Iterable[str]) – a specific set of schedule IDs to remove

Return type:

None

abstract async remove_task(task_id)

Remove the task with the given ID.

Parameters:

task_id (str) – ID of the task to be removed

Raises:

TaskLookupError – if no matching task was found

Return type:

None

abstract async start(exit_stack, event_broker)

Start the event broker.

Parameters:
  • exit_stack (AsyncExitStack) – an asynchronous exit stack which will be processed when the scheduler is shut down

  • event_broker (EventBroker) – the event broker shared between the scheduler, worker (if any) and this data store

Return type:

None

class apscheduler.datastores.memory.MemoryDataStore(tasks=NOTHING, schedules=NOTHING, schedules_by_id=NOTHING, schedules_by_task_id=NOTHING, jobs=NOTHING, jobs_by_id=NOTHING, jobs_by_task_id=NOTHING, job_results=NOTHING, *, lock_expiration_delay=30)

Stores scheduler data in memory, without serializing it.

Can be shared between multiple schedulers within the same event loop.

async acquire_jobs(worker_id, limit=None)

Acquire unclaimed jobs for execution.

This method claims up to the requested number of jobs for the given worker and returns them.

Parameters:
  • worker_id (str) – unique identifier of the worker

  • limit (int | None) – maximum number of jobs to claim and return

Return type:

list[Job]

Returns:

the list of claimed jobs

async acquire_schedules(scheduler_id, limit)

Acquire unclaimed due schedules for processing.

This method claims up to the requested number of schedules for the given scheduler and returns them.

Parameters:
  • scheduler_id – unique identifier of the scheduler

  • limit – maximum number of schedules to claim

Returns:

the list of claimed schedules

async add_job(job)

Add a job to be executed by an eligible worker.

Parameters:

job (Job) – the job object

Return type:

None

async add_schedule(schedule, conflict_policy)

Add or update the given schedule in the data store.

Parameters:
  • schedule (Schedule) – schedule to be added

  • conflict_policy (ConflictPolicy) – policy that determines what to do if there is an existing schedule with the same ID

Return type:

None

async add_task(task)

Add the given task to the store.

If a task with the same ID already exists, it replaces the old one but does NOT affect task accounting (# of running jobs).

Parameters:

task (Task) – the task to be added

Return type:

None

async get_job_result(job_id)

Retrieve the result of a job.

The result is removed from the store after retrieval.

Parameters:

job_id (UUID) – the identifier of the job

Return type:

JobResult | None

Returns:

the result, or None if the result was not found

async get_jobs(ids=None)

Get the list of pending jobs.

Parameters:

ids (Iterable[UUID] | None) – a specific set of job IDs to return, or None to return all jobs

Return type:

list[Job]

Returns:

the list of matching pending jobs, in the order they will be given to workers

async get_next_schedule_run_time()

Return the earliest upcoming run time of all the schedules in the store, or None if there are no active schedules.

Return type:

datetime | None

async get_schedules(ids=None)

Get schedules from the data store.

Parameters:

ids – a specific set of schedule IDs to return, or None to return all schedules

Returns:

the list of matching schedules, in unspecified order

async get_task(task_id)

Get an existing task definition.

Parameters:

task_id (str) – ID of the task to be returned

Return type:

Task

Returns:

the matching task

Raises:

TaskLookupError – if no matching task was found

async get_tasks()

Get all the tasks in this store.

Returns:

a list of tasks, sorted by ID

async release_job(worker_id, task_id, result)

Release the claim on the given job and record the result.

Parameters:
  • worker_id (str) – unique identifier of the worker

  • task_id (str) – the job’s task ID

  • result (JobResult) – the result of the job

Return type:

None

async release_schedules(scheduler_id, schedules)

Release the claims on the given schedules and update them on the store.

Parameters:
  • scheduler_id – unique identifier of the scheduler

  • schedules – the previously claimed schedules

async remove_schedules(ids)

Remove schedules from the data store.

Parameters:

ids (Iterable[str]) – a specific set of schedule IDs to remove

Return type:

None

async remove_task(task_id)

Remove the task with the given ID.

Parameters:

task_id (str) – ID of the task to be removed

Raises:

TaskLookupError – if no matching task was found

Return type:

None

class apscheduler.datastores.sqlalchemy.SQLAlchemyDataStore(engine, schema=None, max_poll_time=1, max_idle_time=60, *, retry_settings=RetrySettings(stop=<tenacity.stop.stop_after_delay object>, wait=<tenacity.wait.wait_exponential object>), lock_expiration_delay=30, serializer=NOTHING, start_from_scratch=False)

Uses a relational database to store data.

When started, this data store creates the appropriate tables on the given database if they’re not already present.

Operations are retried (in accordance to retry_settings) when an operation raises sqlalchemy.OperationalError.

This store has been tested to work with PostgreSQL (asyncpg driver) and MySQL (asyncmy driver).

Parameters:
  • engine (Engine | AsyncEngine) – an asynchronous SQLAlchemy engine

  • schema (str | None) – a database schema name to use, if not the default

async acquire_jobs(worker_id, limit=None)

Acquire unclaimed jobs for execution.

This method claims up to the requested number of jobs for the given worker and returns them.

Parameters:
  • worker_id (str) – unique identifier of the worker

  • limit (int | None) – maximum number of jobs to claim and return

Return type:

list[Job]

Returns:

the list of claimed jobs

async acquire_schedules(scheduler_id, limit)

Acquire unclaimed due schedules for processing.

This method claims up to the requested number of schedules for the given scheduler and returns them.

Parameters:
  • scheduler_id – unique identifier of the scheduler

  • limit – maximum number of schedules to claim

Returns:

the list of claimed schedules

async add_job(job)

Add a job to be executed by an eligible worker.

Parameters:

job (Job) – the job object

Return type:

None

async add_schedule(schedule, conflict_policy)

Add or update the given schedule in the data store.

Parameters:
  • schedule (Schedule) – schedule to be added

  • conflict_policy (ConflictPolicy) – policy that determines what to do if there is an existing schedule with the same ID

Return type:

None

async add_task(task)

Add the given task to the store.

If a task with the same ID already exists, it replaces the old one but does NOT affect task accounting (# of running jobs).

Parameters:

task (Task) – the task to be added

Return type:

None

classmethod from_url(url, **options)

Create a new asynchronous SQLAlchemy data store.

Parameters:
  • url – an SQLAlchemy URL to pass to create_engine() (must use an async dialect like asyncpg or asyncmy)

  • kwargs – keyword arguments to pass to the initializer of this class

Returns:

the newly created data store

async get_job_result(job_id)

Retrieve the result of a job.

The result is removed from the store after retrieval.

Parameters:

job_id (UUID) – the identifier of the job

Return type:

JobResult | None

Returns:

the result, or None if the result was not found

async get_jobs(ids=None)

Get the list of pending jobs.

Parameters:

ids (Iterable[UUID] | None) – a specific set of job IDs to return, or None to return all jobs

Return type:

list[Job]

Returns:

the list of matching pending jobs, in the order they will be given to workers

async get_next_schedule_run_time()

Return the earliest upcoming run time of all the schedules in the store, or None if there are no active schedules.

Return type:

datetime | None

async get_schedules(ids=None)

Get schedules from the data store.

Parameters:

ids – a specific set of schedule IDs to return, or None to return all schedules

Returns:

the list of matching schedules, in unspecified order

async get_task(task_id)

Get an existing task definition.

Parameters:

task_id (str) – ID of the task to be returned

Return type:

Task

Returns:

the matching task

Raises:

TaskLookupError – if no matching task was found

async get_tasks()

Get all the tasks in this store.

Returns:

a list of tasks, sorted by ID

async release_job(worker_id, task_id, result)

Release the claim on the given job and record the result.

Parameters:
  • worker_id (str) – unique identifier of the worker

  • task_id (str) – the job’s task ID

  • result (JobResult) – the result of the job

Return type:

None

async release_schedules(scheduler_id, schedules)

Release the claims on the given schedules and update them on the store.

Parameters:
  • scheduler_id – unique identifier of the scheduler

  • schedules – the previously claimed schedules

async remove_schedules(ids)

Remove schedules from the data store.

Parameters:

ids (Iterable[str]) – a specific set of schedule IDs to remove

Return type:

None

async remove_task(task_id)

Remove the task with the given ID.

Parameters:

task_id (str) – ID of the task to be removed

Raises:

TaskLookupError – if no matching task was found

Return type:

None

async start(exit_stack, event_broker)

Start the event broker.

Parameters:
  • exit_stack (AsyncExitStack) – an asynchronous exit stack which will be processed when the scheduler is shut down

  • event_broker (EventBroker) – the event broker shared between the scheduler, worker (if any) and this data store

Return type:

None

class apscheduler.datastores.mongodb.MongoDBDataStore(client, *, retry_settings=RetrySettings(stop=<tenacity.stop.stop_after_delay object>, wait=<tenacity.wait.wait_exponential object>), lock_expiration_delay=30, serializer=NOTHING, start_from_scratch=False, database='apscheduler')

Uses a MongoDB server to store data.

When started, this data store creates the appropriate indexes on the given database if they’re not already present.

Operations are retried (in accordance to retry_settings) when an operation raises pymongo.errors.ConnectionFailure.

Parameters:
  • client (MongoClient) – a PyMongo client

  • database (str) – name of the database to use

async acquire_jobs(worker_id, limit=None)

Acquire unclaimed jobs for execution.

This method claims up to the requested number of jobs for the given worker and returns them.

Parameters:
  • worker_id (str) – unique identifier of the worker

  • limit (int | None) – maximum number of jobs to claim and return

Return type:

list[Job]

Returns:

the list of claimed jobs

async acquire_schedules(scheduler_id, limit)

Acquire unclaimed due schedules for processing.

This method claims up to the requested number of schedules for the given scheduler and returns them.

Parameters:
  • scheduler_id – unique identifier of the scheduler

  • limit – maximum number of schedules to claim

Returns:

the list of claimed schedules

async add_job(job)

Add a job to be executed by an eligible worker.

Parameters:

job (Job) – the job object

Return type:

None

async add_schedule(schedule, conflict_policy)

Add or update the given schedule in the data store.

Parameters:
  • schedule (Schedule) – schedule to be added

  • conflict_policy (ConflictPolicy) – policy that determines what to do if there is an existing schedule with the same ID

Return type:

None

async add_task(task)

Add the given task to the store.

If a task with the same ID already exists, it replaces the old one but does NOT affect task accounting (# of running jobs).

Parameters:

task (Task) – the task to be added

Return type:

None

async get_job_result(job_id)

Retrieve the result of a job.

The result is removed from the store after retrieval.

Parameters:

job_id (UUID) – the identifier of the job

Return type:

JobResult | None

Returns:

the result, or None if the result was not found

async get_jobs(ids=None)

Get the list of pending jobs.

Parameters:

ids (Iterable[UUID] | None) – a specific set of job IDs to return, or None to return all jobs

Return type:

list[Job]

Returns:

the list of matching pending jobs, in the order they will be given to workers

async get_next_schedule_run_time()

Return the earliest upcoming run time of all the schedules in the store, or None if there are no active schedules.

Return type:

datetime | None

async get_schedules(ids=None)

Get schedules from the data store.

Parameters:

ids – a specific set of schedule IDs to return, or None to return all schedules

Returns:

the list of matching schedules, in unspecified order

async get_task(task_id)

Get an existing task definition.

Parameters:

task_id (str) – ID of the task to be returned

Return type:

Task

Returns:

the matching task

Raises:

TaskLookupError – if no matching task was found

async get_tasks()

Get all the tasks in this store.

Returns:

a list of tasks, sorted by ID

async release_job(worker_id, task_id, result)

Release the claim on the given job and record the result.

Parameters:
  • worker_id (str) – unique identifier of the worker

  • task_id (str) – the job’s task ID

  • result (JobResult) – the result of the job

Return type:

None

async release_schedules(scheduler_id, schedules)

Release the claims on the given schedules and update them on the store.

Parameters:
  • scheduler_id – unique identifier of the scheduler

  • schedules – the previously claimed schedules

async remove_schedules(ids)

Remove schedules from the data store.

Parameters:

ids (Iterable[str]) – a specific set of schedule IDs to remove

Return type:

None

async remove_task(task_id)

Remove the task with the given ID.

Parameters:

task_id (str) – ID of the task to be removed

Raises:

TaskLookupError – if no matching task was found

Return type:

None

async start(exit_stack, event_broker)

Start the event broker.

Parameters:
  • exit_stack (AsyncExitStack) – an asynchronous exit stack which will be processed when the scheduler is shut down

  • event_broker (EventBroker) – the event broker shared between the scheduler, worker (if any) and this data store

Return type:

None

Event brokers

class apscheduler.abc.EventBroker

Interface for objects that can be used to publish notifications to interested subscribers.

abstract async publish(event)

Publish an event.

Return type:

None

abstract async publish_local(event)

Publish an event, but only to local subscribers.

Return type:

None

abstract async start(exit_stack)

Start the event broker.

Parameters:

exit_stack (AsyncExitStack) – an asynchronous exit stack which will be processed when the scheduler is shut down

Return type:

None

abstract subscribe(callback, event_types=None, *, is_async=True, one_shot=False)

Subscribe to events from this event broker.

Parameters:
  • callback – callable to be called with the event object when an event is published

  • event_types – an iterable of concrete Event classes to subscribe to

  • is_asyncTrue if the (synchronous) callback should be called on the event loop thread, False if it should be called in a worker thread. If the callback is a coroutine function, this flag is ignored.

  • one_shot – if True, automatically unsubscribe after the first matching event

class apscheduler.eventbrokers.local.LocalEventBroker

Asynchronous, local event broker.

This event broker only broadcasts within the process it runs in, and is therefore not suitable for multi-node or multiprocess use cases.

Does not serialize events.

async publish(event)

Publish an event.

Return type:

None

class apscheduler.eventbrokers.asyncpg.AsyncpgEventBroker(connection_factory, *, retry_settings=RetrySettings(stop=<tenacity.stop.stop_after_delay object>, wait=<tenacity.wait.wait_exponential object>), serializer=NOTHING, channel='apscheduler', max_idle_time=10)

An asynchronous, asyncpg based event broker that uses a PostgreSQL server to broadcast events using its NOTIFY mechanism.

Parameters:
  • connection_factory (Callable[[], Awaitable[Connection]]) – a callable that creates an asyncpg connection

  • channel (str) – the NOTIFY channel to use

  • max_idle_time (float) – maximum time to let the connection go idle, before sending a SELECT 1 query to prevent a connection timeout

classmethod from_async_sqla_engine(engine, options=None, **kwargs)

Create a new asyncpg event broker from an SQLAlchemy engine.

The engine will only be used to create the appropriate options for asyncpg.connect().

Parameters:
  • engine (AsyncEngine) – an asynchronous SQLAlchemy engine using asyncpg as the driver

  • options (Mapping[str, Any] | None) – extra keyword arguments passed to asyncpg.connect() (will override any automatically generated arguments based on the engine)

  • kwargs (Any) – keyword arguments to pass to the initializer of this class

Return type:

AsyncpgEventBroker

Returns:

the newly created event broker

classmethod from_dsn(dsn, options=None, **kwargs)

Create a new asyncpg event broker from an existing asyncpg connection pool.

Parameters:
  • dsn – data source name, passed as first positional argument to asyncpg.connect()

  • options – keyword arguments passed to asyncpg.connect()

  • kwargs – keyword arguments to pass to the initializer of this class

Returns:

the newly created event broker

async publish(event)

Publish an event.

Return type:

None

async start(exit_stack)

Start the event broker.

Parameters:

exit_stack (AsyncExitStack) – an asynchronous exit stack which will be processed when the scheduler is shut down

Return type:

None

class apscheduler.eventbrokers.mqtt.MQTTEventBroker(client=NOTHING, *, retry_settings=RetrySettings(stop=<tenacity.stop.stop_after_delay object>, wait=<tenacity.wait.wait_exponential object>), serializer=NOTHING, host='localhost', port=1883, topic='apscheduler', subscribe_qos=0, publish_qos=0)

An event broker that uses an MQTT (v3.1 or v5) broker to broadcast events.

Requires the paho-mqtt library to be installed.

Parameters:
  • client (Client) – a paho-mqtt client

  • host (str) – host name or IP address to connect to

  • port (int) – TCP port number to connect to

  • topic (str) – topic on which to send the messages

  • subscribe_qos (int) – MQTT QoS to use for subscribing messages

  • publish_qos (int) – MQTT QoS to use for publishing messages

async publish(event)

Publish an event.

Return type:

None

async start(exit_stack)

Start the event broker.

Parameters:

exit_stack (AsyncExitStack) – an asynchronous exit stack which will be processed when the scheduler is shut down

Return type:

None

class apscheduler.eventbrokers.redis.RedisEventBroker(client, *, retry_settings=RetrySettings(stop=<tenacity.stop.stop_after_delay object>, wait=<tenacity.wait.wait_exponential object>), serializer=NOTHING, channel='apscheduler', stop_check_interval=1)

An event broker that uses a Redis server to broadcast events.

Requires the redis library to be installed.

Parameters:
  • client (Redis | RedisCluster) – an asynchronous Redis client

  • channel (str) – channel on which to send the messages

  • stop_check_interval (float) – interval (in seconds) on which the channel listener should check if it should stop (higher values mean slower reaction time but less CPU use)

classmethod from_url(url, **kwargs)

Create a new event broker from a URL.

Parameters:
  • url (str) – a Redis URL (`redis://...`)

  • kwargs – keyword arguments to pass to the initializer of this class

Return type:

RedisEventBroker

Returns:

the newly created event broker

async publish(event)

Publish an event.

Return type:

None

async start(exit_stack)

Start the event broker.

Parameters:

exit_stack (AsyncExitStack) – an asynchronous exit stack which will be processed when the scheduler is shut down

Return type:

None

Serializers

class apscheduler.abc.Serializer

Interface for classes that implement (de)serialization.

abstract deserialize(serialized)

Restore a previously serialized object from bytestring

Parameters:

serialized (bytes) – a bytestring previously received from serialize()

Return type:

Any

Returns:

a copy of the original object

abstract serialize(obj)

Turn the given object into a bytestring.

Return type:

bytes

Returns:

a bytestring that can be later restored using deserialize()

class apscheduler.serializers.cbor.CBORSerializer(*, type_tag=4664, dump_options=NOTHING, load_options=NOTHING)

Serializes objects using CBOR (RFC 8949).

Can serialize types not normally CBOR serializable, if they implement __getstate__() and __setstate__().

Parameters:
  • type_tag – CBOR tag number for indicating arbitrary serialized object

  • dump_options – keyword arguments passed to cbor2.dumps()

  • load_options – keyword arguments passed to cbor2.loads()

deserialize(serialized)

Restore a previously serialized object from bytestring

Parameters:

serialized (bytes) – a bytestring previously received from serialize()

Returns:

a copy of the original object

serialize(obj)

Turn the given object into a bytestring.

Return type:

bytes

Returns:

a bytestring that can be later restored using deserialize()

class apscheduler.serializers.json.JSONSerializer(*, magic_key='_apscheduler_json', dump_options=NOTHING, load_options=NOTHING)

Serializes objects using JSON.

Can serialize types not normally CBOR serializable, if they implement __getstate__() and __setstate__(). These objects are serialized into dicts that contain the necessary information for deserialization in magic_key.

Parameters:
  • magic_key – name of a specially handled dict key that indicates that a dict contains a serialized instance of an arbitrary type

  • dump_options – keyword arguments passed to json.dumps()

  • load_options – keyword arguments passed to json.loads()

deserialize(serialized)

Restore a previously serialized object from bytestring

Parameters:

serialized (bytes) – a bytestring previously received from serialize()

Returns:

a copy of the original object

serialize(obj)

Turn the given object into a bytestring.

Return type:

bytes

Returns:

a bytestring that can be later restored using deserialize()

class apscheduler.serializers.pickle.PickleSerializer(*, protocol=4)

Uses the pickle module to (de)serialize objects.

As this serialization method is native to Python, it is able to serialize a wide range of types, at the expense of being insecure. Do not use this serializer unless you can fully trust the entire system to not have maliciously injected data. Such data can be made to call arbitrary functions with arbitrary arguments on unpickling.

Parameters:

protocol (int) – the pickle protocol number to use

deserialize(serialized)

Restore a previously serialized object from bytestring

Parameters:

serialized (bytes) – a bytestring previously received from serialize()

Returns:

a copy of the original object

serialize(obj)

Turn the given object into a bytestring.

Return type:

bytes

Returns:

a bytestring that can be later restored using deserialize()

Triggers

class apscheduler.abc.Trigger(*args, **kwds)

Abstract base class that defines the interface that every trigger must implement.

abstract next()

Return the next datetime to fire on.

If no such datetime can be calculated, None is returned. :raises apscheduler.exceptions.MaxIterationsReached:

Return type:

datetime | None

class apscheduler.triggers.date.DateTrigger(run_time)

Triggers once on the given date/time.

Parameters:

run_time (datetime | str | None) – the date/time to run the job at

next()

Return the next datetime to fire on.

If no such datetime can be calculated, None is returned. :raises apscheduler.exceptions.MaxIterationsReached:

Return type:

datetime | None

class apscheduler.triggers.interval.IntervalTrigger(*, weeks=0, days=0, hours=0, minutes=0, seconds=0, microseconds=0, start_time=NOTHING, end_time=None)

Triggers on specified intervals.

The first trigger time is on start_time which is the moment the trigger was created unless specifically overridden. If end_time is specified, the last trigger time will be at or before that time. If no end_time has been given, the trigger will produce new trigger times as long as the resulting datetimes are valid datetimes in Python.

Parameters:
  • weeks (float) – number of weeks to wait

  • days (float) – number of days to wait

  • hours (float) – number of hours to wait

  • minutes (float) – number of minutes to wait

  • seconds (float) – number of seconds to wait

  • microseconds (float) – number of microseconds to wait

  • start_time (datetime | str | None) – first trigger date/time (defaults to current date/time if omitted)

  • end_time (datetime | str | None) – latest possible date/time to trigger on

next()

Return the next datetime to fire on.

If no such datetime can be calculated, None is returned. :raises apscheduler.exceptions.MaxIterationsReached:

Return type:

datetime | None

class apscheduler.triggers.calendarinterval.CalendarIntervalTrigger(*, years=0, months=0, weeks=0, days=0, hour=0, minute=0, second=0, start_date=NOTHING, end_date=None, timezone='local')

Runs the task on specified calendar-based intervals always at the same exact time of day.

When calculating the next date, the years and months parameters are first added to the previous date while keeping the day of the month constant. This is repeated until the resulting date is valid. After that, the weeks and days parameters are added to that date. Finally, the date is combined with the given time (hour, minute, second) to form the final datetime.

This means that if the days or weeks parameters are not used, the task will always be executed on the same day of the month at the same wall clock time, assuming the date and time are valid.

If the resulting datetime is invalid due to a daylight saving forward shift, the date is discarded and the process moves on to the next date. If instead the datetime is ambiguous due to a backward DST shift, the earlier of the two resulting datetimes is used.

If no previous run time is specified when requesting a new run time (like when starting for the first time or resuming after being paused), start_date is used as a reference and the next valid datetime equal to or later than the current time will be returned. Otherwise, the next valid datetime starting from the previous run time is returned, even if it’s in the past.

Warning

Be wary of setting a start date near the end of the month (29. – 31.) if you have months specified in your interval, as this will skip the months when those days do not exist. Likewise, setting the start date on the leap day (February 29th) and having years` defined may cause some years to be skipped.

Users are also discouraged from using a time inside the target timezone’s DST switching period (typically around 2 am) since a date could either be skipped or repeated due to the specified wall clock time either occurring twice or not at all.

Parameters:
  • years (int) – number of years to wait

  • months (int) – number of months to wait

  • weeks (int) – number of weeks to wait

  • days (int) – number of days to wait

  • hour (int) – hour to run the task at

  • minute (int) – minute to run the task at

  • second (int) – second to run the task at

  • start_date (date | str | None) – first date to trigger on (defaults to current date if omitted)

  • end_date (date | str | None) – latest possible date to trigger on

  • timezone (str | tzinfo | None) – time zone to use for calculating the next fire time

next()

Return the next datetime to fire on.

If no such datetime can be calculated, None is returned. :raises apscheduler.exceptions.MaxIterationsReached:

Return type:

datetime | None

class apscheduler.triggers.combining.AndTrigger(triggers, threshold=1, max_iterations=10000)

Fires on times produced by the enclosed triggers whenever the fire times are within the given threshold.

If the produced fire times are not within the given threshold of each other, the trigger(s) that produced the earliest fire time will be asked for their next fire time and the iteration is restarted. If instead all the triggers agree on a fire time, all the triggers are asked for their next fire times and the earliest of the previously produced fire times will be returned.

This trigger will be finished when any of the enclosed trigger has finished.

Parameters:
  • triggers – triggers to combine

  • threshold – maximum time difference between the next fire times of the triggers in order for the earliest of them to be returned from next() (in seconds, or as timedelta)

  • max_iterations – maximum number of iterations of fire time calculations before giving up

next()

Return the next datetime to fire on.

If no such datetime can be calculated, None is returned. :raises apscheduler.exceptions.MaxIterationsReached:

Return type:

datetime | None

class apscheduler.triggers.combining.OrTrigger(triggers)

Fires on every fire time of every trigger in chronological order. If two or more triggers produce the same fire time, it will only be used once.

This trigger will be finished when none of the enclosed triggers can produce any new fire times.

Parameters:

triggers – triggers to combine

next()

Return the next datetime to fire on.

If no such datetime can be calculated, None is returned. :raises apscheduler.exceptions.MaxIterationsReached:

Return type:

datetime | None

class apscheduler.triggers.cron.CronTrigger(*, year=None, month=None, day=None, week=None, day_of_week=None, hour=None, minute=None, second=None, start_time=NOTHING, end_time=None, timezone=NOTHING)

Triggers when current time matches all specified time constraints, similarly to how the UNIX cron scheduler works.

Parameters:
  • year (int | str | None) – 4-digit year

  • month (int | str | None) – month (1-12)

  • day (int | str | None) – day of the (1-31)

  • week (int | str | None) – ISO week (1-53)

  • day_of_week (int | str | None) – number or name of weekday (0-7 or sun,mon,tue,wed,thu,fri,sat, sun)

  • hour (int | str | None) – hour (0-23)

  • minute (int | str | None) – minute (0-59)

  • second (int | str | None) – second (0-59)

  • start_time (datetime | str | None) – earliest possible date/time to trigger on (defaults to current time)

  • end_time (datetime | None) – latest possible date/time to trigger on

  • timezone (str | tzinfo | None) – time zone to use for the date/time calculations (defaults to the local timezone)

Note

The first weekday is always monday.

classmethod from_crontab(expr, timezone='local')

Create a CronTrigger from a standard crontab expression.

See https://en.wikipedia.org/wiki/Cron for more information on the format accepted here.

Parameters:
  • expr (str) – minute, hour, day of month, month, day of week

  • timezone (str | tzinfo) – time zone to use for the date/time calculations (defaults to local timezone if omitted)

Return type:

CronTrigger

next()

Return the next datetime to fire on.

If no such datetime can be calculated, None is returned. :raises apscheduler.exceptions.MaxIterationsReached:

Return type:

datetime | None

Events

class apscheduler.Event(*, timestamp=NOTHING)

Base class for all events.

Variables:

timestamp – the time when the event occurrent

class apscheduler.DataStoreEvent(*, timestamp=NOTHING)

Base class for events originating from a data store.

class apscheduler.TaskAdded(*, timestamp=NOTHING, task_id)

Signals that a new task was added to the store.

Variables:

task_id – ID of the task that was added

class apscheduler.TaskUpdated(*, timestamp=NOTHING, task_id)

Signals that a task was updated in a data store.

Variables:

task_id – ID of the task that was updated

class apscheduler.TaskRemoved(*, timestamp=NOTHING, task_id)

Signals that a task was removed from the store.

Variables:

task_id – ID of the task that was removed

class apscheduler.ScheduleAdded(*, timestamp=NOTHING, schedule_id, next_fire_time)

Signals that a new schedule was added to the store.

Variables:
  • schedule_id – ID of the schedule that was added

  • next_fire_time – the first run time calculated for the schedule

class apscheduler.ScheduleUpdated(*, timestamp=NOTHING, schedule_id, next_fire_time)

Signals that a schedule has been updated in the store.

Variables:
  • schedule_id – ID of the schedule that was updated

  • next_fire_time – the next time the schedule will run

class apscheduler.ScheduleRemoved(*, timestamp=NOTHING, schedule_id)

Signals that a schedule was removed from the store.

Variables:

schedule_id – ID of the schedule that was removed

class apscheduler.JobAdded(*, timestamp=NOTHING, job_id, task_id, schedule_id, tags)

Signals that a new job was added to the store.

Variables:
  • job_id – ID of the job that was added

  • task_id – ID of the task the job would run

  • schedule_id – ID of the schedule the job was created from

  • tags – the set of tags collected from the associated task and schedule

class apscheduler.JobRemoved(*, timestamp=NOTHING, job_id)

Signals that a job was removed from the store.

Variables:

job_id – ID of the job that was removed

class apscheduler.ScheduleDeserializationFailed(*, timestamp=NOTHING, schedule_id, exception)

Signals that the deserialization of a schedule has failed.

Variables:
  • schedule_id – ID of the schedule that failed to deserialize

  • exception – the exception that was raised during deserialization

class apscheduler.JobDeserializationFailed(*, timestamp=NOTHING, job_id, exception)

Signals that the deserialization of a job has failed.

Variables:
  • job_id – ID of the job that failed to deserialize

  • exception – the exception that was raised during deserialization

class apscheduler.SchedulerEvent(*, timestamp=NOTHING)

Base class for events originating from a scheduler.

class apscheduler.SchedulerStarted(*, timestamp=NOTHING)
class apscheduler.SchedulerStopped(*, timestamp=NOTHING, exception=None)

Signals that a scheduler has stopped.

Variables:

exception – the exception that caused the scheduler to stop, if any

class apscheduler.JobAcquired(*, timestamp=NOTHING, job_id, worker_id)

Signals that a worker has acquired a job for processing.

Parameters:
  • job_id (UUID | str) – the ID of the job that was acquired

  • worker_id (str) – the ID of the worker that acquired the job

class apscheduler.JobReleased(*, timestamp=NOTHING, job_id, worker_id, outcome, exception_type=None, exception_message=None, exception_traceback=None)

Signals that a worker has finished processing of a job.

Parameters:
  • job_id (UUID | str) – the ID of the job that was released

  • worker_id (str) – the ID of the worker that released the job

  • outcome (TEnum | str) – the outcome of the job

  • exception_type (str | None) – the fully qualified name of the exception if outcome is JobOutcome.error

  • exception_message (str | None) – the result of str(exception) if outcome is JobOutcome.error

  • exception_traceback (list[str] | None) – the traceback lines from the exception if outcome is JobOutcome.error

Enumerated types

class apscheduler.RunState(value)

Used to track the running state of schedulers and workers.

Values:

  • starting: not running yet, but in the process of starting

  • started: running

  • stopping: still running but in the process of shutting down

  • stopped: not running

class apscheduler.JobOutcome(value)

Used to indicate how the execution of a job ended.

Values:

  • success: the job completed successfully

  • error: the job raised an exception

  • missed_start_deadline: the job’s execution was delayed enough for it to miss its deadline

  • cancelled: the job’s execution was cancelled

class apscheduler.ConflictPolicy(value)

Used to indicate what to do when trying to add a schedule whose ID conflicts with an existing schedule.

Values:

  • replace: replace the existing schedule with a new one

  • do_nothing: keep the existing schedule as-is and drop the new schedule

  • exception: raise an exception if a conflict is detected

class apscheduler.CoalescePolicy(value)

Used to indicate how to queue jobs for a schedule that has accumulated multiple run times since the last scheduler iteration.

Values:

  • earliest: run once, with the earliest fire time

  • latest: run once, with the latest fire time

  • all: submit one job for every accumulated fire time

Context variables

See the contextvars module for information on how to work with context variables.

apscheduler.current_scheduler the current scheduler: ContextVar[Union[Scheduler, AsyncScheduler]]
apscheduler.current_worker the current scheduler: ContextVar[Union[Worker, AsyncWorker]]
apscheduler.current_job information on the job being currently run: ContextVar[JobInfo]

Exceptions

exception apscheduler.TaskLookupError(task_id)

Raised by a data store when it cannot find the requested task.

exception apscheduler.ScheduleLookupError(schedule_id)

Raised by a scheduler when it cannot find the requested schedule.

exception apscheduler.JobLookupError(job_id)

Raised when the job store cannot find a job for update or removal.

exception apscheduler.JobResultNotReady(job_id)

Raised by get_job_result() if the job result is not ready.

exception apscheduler.JobCancelled

Raised by get_job_result() if the job was cancelled.

exception apscheduler.JobDeadlineMissed

Raised by get_job_result() if the job failed to start within the allotted time.

exception apscheduler.ConflictingIdError(schedule_id)

Raised when trying to add a schedule to a store that already contains a schedule by that ID, and the conflict policy of exception is used.

exception apscheduler.SerializationError

Raised when a serializer fails to serialize the given object.

exception apscheduler.DeserializationError

Raised when a serializer fails to deserialize the given object.

exception apscheduler.MaxIterationsReached

Raised when a trigger has reached its maximum number of allowed computation iterations when trying to calculate the next fire time.