graphile-worker

graphile-worker

Job queue for PostgreSQL running on Node.js - allows you to run jobs (e.g. sending emails, performing calculations, generating PDFs, etc) "in the background" so that your HTTP response is not held up. Can be used with any PostgreSQL-backed application. Pairs beautifully with PostGraphile.

Quickstart

In your existing Node.js project:

Add the worker to your project:

yarn add graphile-worker

Create tasks:

Create a tasks folder, and place in it JS files containing your task specs. The names of these files will be the task identifiers, e.g. hello below:

// tasks/hello.js
module.exports = async ({ name }) => {
  console.log(`Hello, ${name}`);
};

Run the worker

(Make sure you're in the folder that contains the tasks folder.)

npx graphile-worker -c "my_db"
# or, if you have a remote database, something like:
# npx graphile-worker -c "postgres://user:pass@host:port/db?ssl=1"
# or, if you prefer envvars
# DATABASE_URL="..." npx graphile-worker

(Note: npx runs the local copy of an npm module if it is installed, when you're ready, switch to using the package.json "scripts" entry instead.)

Schedule a job:

Connect to your database and run the following SQL:

SELECT graphile_worker.add_job('hello', json_build_object('name', 'Bobby Tables'));

Success!

You should see the worker output Hello, Bobby Tables. Gosh, that was fast!

Features

• Standalone and embedded modes
• Easy to test with (including runAllJobs util)
• Low latency (~2ms from task schedule to execution, uses LISTEN/NOTIFY to be informed of jobs as they're inserted)
• High performance (~700 jobs per second, uses SKIP LOCKED to find jobs to execute, resulting in faster fetches)
• Small tasks (uses explicit task names / payloads resulting in minimal serialisation/deserialisation overhead)
• Parallel by default
• Adding jobs to same named queue runs them in series
• Automatically re-attempts failed jobs with exponential back-off
• Customisable retry count (default: 25 attempts over ~3 days)
• Open source
• Executes tasks written in Node.js (can call out to any other language or networked service)
• Modern JS with async/await
• Watch mode for development (experimental - iterate your jobs without restarting worker)

Crowd-funded open-source software

In exchange for the incredible freedom we give in how this software can be used (thanks to the permissive MIT license) we ask the individuals and businesses that use it to sponsor ongoing maintenance and development via sponsorship.

Status

Seems stable and has good test suite, but needs real-world testing before it can be deemed production ready. Reach out on GitHub issues or Discord chat if you'd like to help with this: http://discord.gg/graphile

Requirements

PostgreSQL 10+ and Node 10+.

If your database doesn't already include the pgcrypto and uuid-ossp extensions we'll automatically install them into the public schema for you. If you have them installed in a different schema (unlikely) you may face issues.

Installation

yarn add graphile-worker

Running:

graphile-worker manages it's own database schema (graphile_worker) just point graphile-worker at your database and we handle our own migrations:

npx graphile-worker -c "postgres://localhost/mydb"

(npx looks for the graphile-worker binary locally; it's often better to use the "scripts" entry in package.json instead.)

Creating task executors

A task executor is a simple async JS function which receives as input the job payload and a collection of helpers. It does the work and then returns. If it returns then the job is deemed a success and is deleted from the queue. If it throws an error then the job is deemed a failure and the task is rescheduled using an exponential backoff algorithm.

IMPORTANT: your jobs should wait for all asynchronous work to be completed *before returning, otherwise we might mistakenly think they were successful.

IMPORTANT: we automatically retry the job if it fails, so it's often sensible to split large jobs into smaller jobs, this also allows them to run in parallel resulting in faster execution. This is particularly important for tasks that are not idempotent (i.e. running them a second time will have extra side effects) - for example sending emails.

Tasks are created in the tasks folder in the directory from which you run graphile-worker; the name of the file (less the .js suffix) is used as the task identifier. Currently only .js files that can be directly loaded by Node.js are supported; if you are using Babel, TypeScript or similar you will need to compile your tasks into the tasks folder.

current directory
├── package.json
├── node_modules
└── tasks
    ├── task_1.js
    └── task_2.js

// tasks/task_1.js
module.exports = async payload => {
  await doMyLogicWith(payload);
};

// tasks/task_2.js
module.exports = async (payload, { debug }) => {
  // async is optional, but best practice
  debug(`Received ${JSON.stringify(payload)}`);
};

Each task function is passed two arguments:

• payload - the payload you passed when calling add_job
• helpers - an object containing:
  • debug - a helpful debug instance scoped to the name of the task (use the DEBUG envvar to expose)
  • job - the whole job (including uuid, attempts, etc) - you shouldn't need this
  • withPgClient - a helper to use to get a database client

withPgClient example:

const {
  rows: [row]
} = await withPgClient(pgClient => pgClient.query("select 1 as one"));

Scheduling jobs

You can schedule jobs directly in the database, e.g. from a trigger or function, or by calling SQL from your application code. You do this using the graphile_worker.add_job function. (We'll add a JS helper for this soon...)

add_job accepts the following parameters (in this order):

• identifier - the only required field, indicates the name of the task executor to run (omit the .js suffix!)
• payload - a JSON object with information to tell the task executor what to do (defaults to an empty object)
• queue_name - if you want certain tasks to run one at a time, add them to the same named queue (defaults to a random value)
• run_at - a timestamp after which to run the job; defaults to now.
• max_attempts - if this task fails, how many times should we retry it? Default: 25.

Typically you'll want to set the identifier and payload:

SELECT graphile_worker.add_job(
  'send_email',
  json_build_object(
    'to', '[e-mail]',
    'subject', 'graphile-worker test'
  )
);

You can skip parameters you don't need by using PostgreSQL's named parameter support:

SELECT graphile_worker.add_job('reminder', run_at := NOW() + INTERVAL '2 days');

Uninstallation

To delete the worker code and all the tasks from your database, just run this one SQL statement:

DROP SCHEMA graphile_worker CASCADE;

Performance

graphile-worker is not intended to replace extremely high performance dedicated job queues, it's intended to be a very easy way to get a job queue up and running with Node.js and PostgreSQL. But this doesn't mean it's a slouch by any means - it achieves an average latency from triggering a job in one process to executing it in another of just 2ms, and each worker can handle up to 731 jobs per second on modest hardware (2011 iMac).

graphile-worker is horizontally scalable. Each instance has a customisable worker pool, this pool defaults to size 1 (only one job at a time on this worker) but depending on the nature of your tasks (i.e. assuming they're not compute-heavy) you will likely want to set this higher to benefit from Node.js' concurrency. If your tasks are compute heavy you may still wish to set it higher and then using Node's child_process (or Node v11's worker_threads) to share the compute load over multiple cores without significantly impacting the main worker's runloop.

To test performance you can run yarn perfTest. This reveals that on a 2011 iMac running both the worker and the database (and a bunch of other stuff) starting the command, checking for jobs, and exiting takes about 0.40s and running 20,000 trivial queued jobs across a single worker pool of size 1 takes 27.35s (~731 jobs per second). Latencies are also measured, from before the call to queue the job is fired until when the job is actually executed. These latencies ranged from 1.39ms to 19.66ms with an average of 1.90ms.

Exponential backoff

We currently use the formula exp(least(10, attempt)) to determine the delays between attempts (the job must fail before the next attempt is scheduled, so the total time elapsed may be greater depending on how long the job runs for before it fails). This seems to handle temporary issues well, after ~4 hours attempts will be made every ~6 hours until the maximum number of attempts is achieved. The specific delays can be seen below:

select
  attempt,
  exp(least(10, attempt)) * interval '1 second' as delay,
  sum(exp(least(10, attempt)) * interval '1 second') over (order by attempt asc) total_delay
from generate_series(1, 24) as attempt;

 attempt |      delay      |   total_delay
---------+-----------------+-----------------
       1 | 00:00:02.718282 | 00:00:02.718282
       2 | 00:00:07.389056 | 00:00:10.107338
       3 | 00:00:20.085537 | 00:00:30.192875
       4 | 00:00:54.598150 | 00:01:24.791025
       5 | 00:02:28.413159 | 00:03:53.204184
       6 | 00:06:43.428793 | 00:10:36.632977
       7 | 00:18:16.633158 | 00:28:53.266135
       8 | 00:49:40.957987 | 01:18:34.224122
       9 | 02:15:03.083928 | 03:33:37.308050
      10 | 06:07:06.465795 | 09:40:43.773845
      11 | 06:07:06.465795 | 15:47:50.239640
      12 | 06:07:06.465795 | 21:54:56.705435
      13 | 06:07:06.465795 | 28:02:03.171230
      14 | 06:07:06.465795 | 34:09:09.637025
      15 | 06:07:06.465795 | 40:16:16.102820
      16 | 06:07:06.465795 | 46:23:22.568615
      17 | 06:07:06.465795 | 52:30:29.034410
      18 | 06:07:06.465795 | 58:37:35.500205
      19 | 06:07:06.465795 | 64:44:41.966000
      20 | 06:07:06.465795 | 70:51:48.431795
      21 | 06:07:06.465795 | 76:58:54.897590
      22 | 06:07:06.465795 | 83:06:01.363385
      23 | 06:07:06.465795 | 89:13:07.829180
      24 | 06:07:06.465795 | 95:20:14.294975

Development

yarn
yarn watch

In another terminal:

createdb graphile_worker_test
yarn test