Each instance of a Background Task is associated with a Task object that can be used to get the task’s current status, and any data the task has passed to the main program.
The Task object can also be used to control the running of the task.
From within the task, modify
anvil.server.task_state to share data with the main program.
The main program gets a Task object as the return value of
anvil.server.launch_background_task. This object
get_state() method that returns the
anvil.server.task_state is an empty
dict, so you can immediately set key/value pairs:
anvil.server.task_state['progress'] = 42.
The Task can also raise exceptions; the main program can access these using the
get_error methods of the Task object.
@anvil.server.callable def launch_my_task(): task = anvil.server.launch_background_task('train_my_network', 200) # Output some state from the task print(task.get_state()['false_positive_rate']) # Is the task complete yet? if task.is_completed(): return # Store the ID of the task so we can see it later app_tables.tasks.add_row( when=datetime.now(), task_id=task.get_id(), ) # If the task raised an exception, re-raise it here task.get_error() # Otherwise, kill the task task.kill() # Check the termination status if task.get_termination_status() == "killed": print("Yes, the task was killed!") # Finally, return the task to the client return task
The Task object reports on task status and gives you control over the task.
There are three ways to get the Task object for a particular task:
- When launching a task. It is the return value of
- Later on, using
anvil.server.get_background_task(task_id). You need to know the
task_id, which you can get using
task.get_id(). You may wish to store this ID in a Data Table or elsewhere.
anvil.server.list_background_tasks(). See “Listing Background Tasks from code”, below.
The Task object’s methods are:
get_state(): returns an object that you can modify from within the task (by modifying
anvil.server.task_stateinside the task).
get_id(): returns the ID of a task, which is a string.
get_task_name(): returns the name of the task. Usually the name of the function decorated with
get_start_time(): returns a
DateTimerepresenting the time this task was started.
get_return_value(): returns the return value of the task, or
Noneif the task has not yet returned.
get_error(): re-throws whatever exception caused the task to terminate.
kill(): stops the task and makes its status
"failed". This can only be executed on the server side.
Noneif the task is still running
"completed"if the task completed successfully
"failed"if the task terminated with an exception
"killed"if the task was explicitly killed with
task.kill(), or by clicking “Kill” in the development environment
"missing"if an infrastructure failure caused us to lose track of the task.
Trueif the task is still running. Otherwise returns
False, regardless of how the task ended (completed, killed, failed or missing).
Trueif the task completed successfully,
Falseif the task is still running. If the task has failed or been killed, it raises whatever caused the task to terminate.
When you run a
For those on the Free Plan, the timeout for background tasks is 30 seconds.
Separation of processes
Background Tasks are run in a separate Python process to your server functions. This means that they’ll never share global variables or
anvil.server.session with your server calls.
The reason we run Background Tasks in a separate process is to do with timeouts: If a server call takes too long, we need to kill it rather than letting it spin forever. Background tasks, however, can run for a long time. If we ran them in the same Python interpreter as your standard server functions, we would have to kill your background tasks whenever a server call timed out.
If the Task needs access to information from
anvil.server.session, then it is the caller’s responsibility to furnish it, via additional parameters passed to
anvil.server.sessionobject is not serialisable. If you need to share session data with a background task, pass only the arguments you need to pass.
Task object methods
The data in
anvil.server.task_state is specific to the task - that is, a background task’s own
anvil.server.task_state is read/write within the Task, and read-only everywhere else, including in the code of other Tasks.
The Uplink doesn’t form part of our managed infrastructure, meaning that if an Uplink connection is disconnected, we lose track of any background tasks.
Because we can’t guarantee to keep track of all background tasks launched via the Uplink, the
get_state() method always returns
None if a task is running on the Uplink.
You will still have access to more general methods relating to the Task object, such as