Aborting Tasks

If a task run takes too long, the Manager will attempt to abort it, and potentially retry it.

Defining timeouts

To define timeouts, every task run has a maxDuration configuration property. This will have been provided when the task run was scheduled, or will have been inherited from the task, the task's category, or the default (1000ms).

If the task run exceeds this duration, it will be considered 'aborted'. However, you will need to write your task in such a way that this abort signal is checked for and acted upon. (JavaScript has no way of aborting arbitrary code elsewhere in the event loop). This is done as follows.

Checking for abort signals

As well as its string parameter, every Task function is passed a signal object as its second parameter. This is an instance of the AbortSignal class, which is part of the standard JavaScript API. You can use this signal to check if a request to abort the task run has been made by the Manager, and to stop execution if it has.

This example creates a task that runs indefinitely, checking for an abort signal every 10ms, and printing out when it gets aborted:

import {createManager} from 'tinytick';

const manager = createManager();
manager.start();

manager.setTask('loops', async (_, signal) => {
  while (!signal.aborted) {
    await pause(10);
  }
  console.log(`Aborted!`);
});

// A convenience function for waiting asynchronously:
const pause = async (ms) => new Promise((resolve) => setTimeout(resolve, ms));

Trying it out

Let's schedule this task to run as soon as possible with a maxDuration of 300ms:

manager.scheduleTaskRun('loops', '', 0, {maxDuration: 300});

After 100ms (the Manager tick interval), the task run will be running:

// ... wait 100ms for task run to start
console.log(manager.getRunningTaskRunIds().length);
// -> 1

After a further 400ms, we can be sure the task run will have been aborted:

// ... wait 400ms for task run to timeout and get aborted
// -> 'Aborted!'

console.log(manager.getRunningTaskRunIds().length);
// -> 0

Note that, as well as checking the aborted property of the signal, you can also listen for the abort event on the signal object using the its addEventListener method.

It's also worth noting that the AbortSignal object passed to a Task function is the same as that used in the fetch platform function, and so you can easily make your network-based tasks abortable by simply passing it on:

manager.setTask('ping', async (url, signal) => await fetch(url, {signal}));

Summary

That's a review of how TinyTick handles aborting tasks. Next, we'll look at how to retry tasks that fail - in the Retries And Backoff guide.