titarenko/joba

Simple task (job) scheduler (manager).

joba

Simple job (task) scheduler (manager).

Installation

npm install joba --save

Depending on your choice you should also install bus and persistence dependencies (for example, joba-worque, joba-pg).

Example

Task manager module.

var Joba = require('joba');

var manager = new Joba({
	persistence: {
		name: 'pg',
		connection: {
			uri: 'postgresql://user:password@server/db',
			table: 'tasks'
		}
	},
	bus: {
		name: 'worque',
		connection: {
			host: 'server',
			login: 'user',
			password: 'password'
		}
	}
});

module.exports = manager;

Scheduler app.

var manager = require('./manager');

manager.schedule('* */2 * * *', 'scan web');

Worker app.

var manager = require('./manager');

manager.handle('scan web', scanWeb, true);

function scanWeb () {
	var profitPromise;
	// do something
	return profitPromise;
}

It's recommended to split your task handling process into 2 parts: scheduling and processing and run them into separate applications. It will allow you to restart worker app on failure without loosing the schedule or task being processed while failure has happened.

API

Joba(options)

Constructs manager instance. Accepts 2 options: bus and persistence configuration (see example).

schedule(cronTime, name[, params])

Schedules task execution.

start(name[, params])

Starts immediate task execution.

handle(name, handler[, exitOnFailure][, logParams])

Deprecated (use overload with options as last param).

Specifies task handler. If exitOnFailure is true, then app will exit with code 75 (temporary failure), allowing you to implement restart-on-error management policy for worker app. If logParams is true then params will be saved in the corresponding work log record, if logParams is omitted, then logging will happen only on error, otherwise (if logParams === false) no logging will happen at all.

handle(name, handler[, options])

Specifies task handler.

Options:

name	purpose
exitOnFailure	if true, then app will exit with code 75 (temporary failure), allowing you to implement restart-on-error management policy for worker app
logParams	if true then params will be saved in the corresponding work log record, if omitted, then logging will happen only on error, otherwise (if false) no logging will happen at all
concurrency	maximal number of tasks handled simultaneously

pipe(source, destination)

Passes output of one task to input(s) of another task(s).

If destination is string:

source -> destination

otherwise, if destination is object and destination.opcode is map:

source[0] -> destination.names[0]
source[1] -> destination.names[0]
source[2] -> destination.names[0]
...

and if destination.opcode is spread:

source[0] -> destination.names[0]
source[1] -> destination.names[1]
source[2] -> destination.names[2]
...

License

BSD