Job NPM | npm.io

Job

What is Job?

Job is a delayed jobs scheduler inspired by the Ruby project delayed_job and written for SpanDeX.io.

Where before you might write mailer.registrationEmail('userID'), you can now write mailer.delay.registrationEmail(5*1000, 'medium', 'userID') to delay the email by 5 minutes.

Installation

Add to package.json and install via npm:

npm install job

And wherever you bootstrap your application, you need to require('job') and initialize it:

require('job').startJobsRunner({ kue: { username: 'example', password: 'password', port: 2442 }})

Optionally, you may provide the path of an initialization script that will bootstrap the jobs runner, since all delayed jobs are run in a separate process. You may need to initialize a database connection, for example. In the future I would like to better support Architect plugins to eliminate the need for a dedicated startup script.

require('job').startJobsRunner({ initScript: './init_job_worker.js', kue: { username: 'example', password: 'password', port: 2442 }})

And the initialization script should be in the form:

module.exports = function initJobsWorker (done) {
  // Initialize database connections, etc
  done();
};

Then, in any modules where you want to have a delayed job:

exports.registationEmail = function (userID) {
  model.Users.findById(userID).exec(function (err, user) {
    // ... send out email
  });
};

becomes:

exports.registationEmail = function (userID) {
  model.Users.findById(userID).exec(function (err, user) {
    // ... send out email
  });
};
exports.registationEmail.delayable = true;

require('job').addHelper(module);

That's it! Note that your helper method must take database IDs, not objects; generally, anything passed into the helper method will first be serialized into Redis, and furthermore will be accessed in a different process. Thus, things like Mongoose objects or anything that requires lookup in an in-memory cache are out.

Delay the method by calling:

lib.delay.registrationEmail(ms, 'priority', args)

where ms is the number of milliseconds to delay by, and priority is 'low', 'medium', 'high', or 'critical'.

As a job queue server

Job can also be used as a simple queue server. For example, if you want to perform an operation ASAP but only want to perform X operations simultaneously:

exports.method = function () { ... }
exports.method.delayable = true;
exports.method.concurrentProcesses = 5;

Then, you can call lib.delay.method(0, 'critical', ...).

By default, concurrentProcesses is 1.

Kue server

Job uses a Kue server to process jobs. The Kue username, password, and port you provide can be used to monitor the status of delayed jobs.

For example, if SpanDeX.io had supplied a port of 1234, we could go to SpanDeX.io:1234, type in the username and password, and monitor the health of all delayed jobs.