Momo scheduler

Momo is a light-weight, easy-to-use scheduler that persists jobs in a MongoDB and supports interval-based scheduling as well as cron jobs. In essence, it provides an easy way to tell your application to either "run that job every five minutes" or "run that job at 9 AM every weekday".

Features

allows concurrent jobs
allows immediate jobs
supports long-running jobs
allows error handling
supports cluster mode (e.g. several instances using the same job database)
- only one schedule per name will attempt to schedule jobs
- if an active schedule stops running (e.g. the application instance was stopped), another schedule with the same name (e.g. running on a different instance) will take over

Installation

Install via npm:

npm install @tngtech/momo-scheduler

or yarn:

yarn add @tngtech/momo-scheduler

You're all set!

Usage

import { MongoSchedule, MomoJob, MomoJobBuilder, MomoErrorEvent, MomoEvent } from 'momo-scheduler';
import { MongoMemoryServer } from 'mongodb-memory-server';
import { MongoScheduleBuilder } from './MongoScheduleBuilder';

const mongo = await MongoMemoryServer.create();
const mongoSchedule = await new MongoScheduleBuilder()
  .withConnection({
    url: mongo.getUri(),
    collectionsPrefix: 'momo',
    scheduleName: 'MySchedule',
    pingIntervalMs: 10_000,
  })
  .build();

const intervalJob: MomoJob = new MomoJobBuilder()
  .withName('interval job')
  .withConcurrency(2)
  .withMaxRunning(3)
  .withSchedule('5 seconds')
  .withParameters({ foo: 'bar' })
  .withHandler(() => console.log('This is a momo job that runs twice every five seconds!'))
  .build();
await mongoSchedule.define(intervalJob);

const cronJob: MomoJob = new MomoJobBuilder()
  .withName('cron job')
  .withConcurrency(1)
  .withMaxRunning(1)
  .withCronSchedule('0 0 * * 1-5')
  .withParameters({ foo: { bar: 'baz' } })
  .withHandler(() => console.log('This is a momo job that runs every weekday at midnight!'))
  .build();
await mongoSchedule.define(cronJob);

const cronJobWithoutParameters: MomoJob = new MomoJobBuilder()
  .withName('cron job without parameters')
  .withConcurrency(1)
  .withMaxRunning(3)
  .withCronSchedule('0 0 * * 1-3')
  .withHandler(() => console.log('This is a momo job that runs on Monday, Tuesday and Wednesday!'))
  .build();
await mongoSchedule.define(cronJobWithoutParameters);

// optional: listen to error and debug events
mongoSchedule.on('error', (error: MomoErrorEvent) => {
  /* handle error */
});
mongoSchedule.on('debug', (debug: MomoEvent) => {
  /* ... */
});

await mongoSchedule.start();

// all your application code...

// gracefully disconnect the scheduler
await mongoSchedule.disconnect();

MomoJob

You can instantiate a momo job using the MomoJobBuilder class. It provides the following setter methods:

setter	parameter	mandatory	default value	description
withName	`name: string`	yes		The name of the job. Used as a unique identifier.
withSchedule	`interval: number \| string`, `firstRunAfter: number \| string` (default: `0`)	no		Either this setter or `withCronSchedule()` must be called. Specifies the schedule for starting the job. Time intervals can be given in human-readable formats (like '1 minute', 'ten days' or 'twenty-one days and 2 hours' ) or in milliseconds. Check the documentation of human-interval for details. If the job never ran before, the job will run after `firstRunAfter` for the first time. Just like the `interval`, `firstRunAfter` can also be given in a human-readable format or as milliseconds.
withCronSchedule	`cronSchedule: string`	no		Either this setter or `withCronSchedule()` must be called. Specifies the cron schedule according to which the job will run. See https://www.npmjs.com/package/cron#cron-ranges for the allowed cron ranges.
withConcurrency	`concurrency: number`	no	1	How many instances of a job are started at a time.
withMaxRunning	`maxRunning: number`	no	0	Maximum number of job executions that is allowed at a time. Set to 0 for no max. The schedule will trigger no more job executions if maxRunning is reached.
withHandler	`handler: function`	yes		The function to execute.
withParameters	`jobParameters: JobParameters`	no		The parameters the `handler` function is called with.

MongoSchedule

You can instantiate a mongo schedule using the MongoScheduleBuilder class. It provides the following setter methods:

setter	parameter	mandatory	default value	description
withJob	`job: MomoJob`	yes		Adds a job to the schedule.
withConnection	`options: MomoOptions`	yes		The connection options of the schedule.

The schedule offers the following methods to create and run jobs:

function	parameters	description
connect	`options: MomoConnectionOptions`	Creates a new MongoSchedule connected to your database. See below for the available options.
disconnect		Cancels all jobs and disconnects the MongoSchedule from the database.
define	`job: MomoJob`	Creates a new job on the schedule or updates an existing one with the same name. If the job existed before, it will be stopped!
start		Starts jobs that are on the schedule.
stop		Stops jobs, but does not remove them from either the schedule or the database.
cancel		Stops and removes jobs from the schedule, does not remove them from the database.
remove		Stops and removes jobs from both the schedule and the database.
count	`onlyStarted: boolean` (optional)	Returns the number of jobs on the schedule. Only started jobs are counted if parameter is set to true.
list		Returns descriptions of all jobs on the schedule. (See below)
check	`name: string`	Returns execution information `{ lastStarted, lastFinished, lastResult }` of the job with the provided name from the database. This also works if the job is not on the schedule.
clear		Removes all jobs from the database. This also removes jobs that are not on this schedule, but were defined by other schedules. However, does NOT stop job executions - this will cause currently running jobs to fail. Consider using stop/cancel/remove methods instead!
get	`name: string`	Returns a description of the job. Returns undefined if no job with the provided name is defined on this schedule.
run	`name: string`, `parameters: JobParameters` (default: `undefined`), `delay: number` (default: `0`)	Runs the job with the provided name once, independently from the schedule, after the specified delay. If no parameters are provided, the job will be called with no parameters. Note that `maxRunning` is respected, ie. the execution is skipped if the job is already running `maxRunning` times.
on	`event: 'debug' \| 'error'`, `listener: function`	Define a callback for debug or error events. (See also below.)

The job descriptions returned by the list and get functions of the MongoSchedule consist of some properties of the MomoJob (see above) as stored in the database, namely the name, the schedule (human-readable or cron), the concurrency and maxRunning and the parameters (if any). If the job was already started, it also contains the schedulerStatus, which reports the number of currently running executions of the job and the schedule the job was started with. This might differ from the top-level schedule as running jobs do not automatically pick up changes to the schedule from the database.

MomoConnectionOptions

property	type	mandatory	default	description
url	`string`	yes		The connection string of your database.
scheduleName	`string`	yes		Only one schedule per name can be active at a time. If multiple instances of your application define a schedule with the same name, only one at a time will actually run jobs.
collectionsPrefix	`string`	no	no prefix	A prefix for all collections created by Momo.
pingIntervalMs	number	no	`60_000`	The keep alive ping interval of the schedule, in milliseconds. After twice the amount of time has elapsed without a ping of your Momo instance, other instances may take over. You might want to reduce this if you have jobs running on short intervals.
mongoClientOptions	`MongoClientOptions`	no		Options for the connection to the MongoDB client as specified by the MongoDB API. Useful for providing configuration options that are not available via the connection string (url).

Reacting to events

The MongoSchedule is an EventEmitter, emitting 'debug' and 'error' events. You can define callbacks to handle them:

mongoSchedule.on('error', ({ error, data, message, type }: MomoErrorEvent) => {
  console.log(`An error of type ${type} occurred`, { error, data, message });
});

mongoSchedule.on('debug', ({ data, message }: MomoEvent) => {
  console.log('A debug event occurred', { data, message });
});

MomoEvent and MomoErrorEvent

event	property	type	description
both	message	`string`	Some information about the event that occurred.
both	data (optional)	`{ name?: string; ... }`	Contains additional information like the name of the affected job.
error	type	`MomoErrorType`	one of: `'defining job failed'`, `'scheduling job failed'`, `'executing job failed'`, `'stopping job failed'`, `'an internal error occurred'`
error	error (optional)	`Error`	The root cause of the error.

Job Examples

import { MomoJobBuilder } from './MomoJobBuilder';

const example1 = new MomoJobBuilder()
  .withName('example 1')
  .withInterval('5 minutes')
  .withHandler(() => console.log('This is momo'))
  .build();

const example2 = new MomoJobBuilder()
  .withName('example 2')
  .withInterval('5 minutes', 60 * 1000) // first run after one minute
  .withHandler(() => console.log('This is momo'))
  .build();

const example3 = new MomoJobBuilder()
  .withName('example 3')
  .withInterval('5 minutes', '4 minutes') // first run after four minutes
  .withHandler(() => console.log('This is momo'))
  .build();

const example4 = new MomoJobBuilder()
  .withName('example 4')
  .withCronSchedule('0 0 * * 1-5') // every weekday at midnight
  .withHandler(() => console.log('This is momo'))
  .build();

Assume it is 12:00 AM when the MongoSchedule with these four example jobs is started.

example 1 will be run immediately, at 12:00, and then every five minutes. Adding firstRunAfter: 0 explicitly is equivalent as this is the default value.
example 2 will be run after 1 minute (the configured firstRunAfter), at 12:01, and then every five minutes (next is 12:06).
example 3 will be run after 4 minutes (the configured firstRunAfter), at 12:04, and then every five minutes (next is 12:09).
example 4 will be run at midnight every weekday

Now assume the MongoSchedule is stopped at 12:02 and then immediately started again.

example 1 will be run 5 minutes (the configured interval) after the last execution, at 12:05. The job is NOT run immediately because it already ran before.
example 2 will be run 5 minutes (the configured interval) after the last execution, at 12:06. The job is NOT run after 1 minute (the configured firstRunAfter) because it already ran before.
example 3 will be run 4 minutes after the start (because of the configured firstRunAfter), at 12:06, because it never ran before, and then every five minutes (next is 12:11).
example 4 will be run at midnight every weekday

If your application is not running while a cron job should start, the missed run will not be started when your application resumes. This is the case for interval jobs, though - if the time of the next run has already elapsed after the application restarted, the job will be run immediately.

The benefit of cron jobs is that you can control the exact time of execution. They are also not restricted to a constant interval.

Supported Node Versions

momo-scheduler is currently tested on the node LTS versions 18 and 20.

License

This project is open source and licensed under Apache 2.0.

Licenses and Copyrights of dependencies can be found in THIRD_PARTY_NOTICE for the latest release.

For contribution guidelines see CONTRIBUTING.md.

TNG/momo-scheduler