An mgFx connector that uses Redis as a transport provider

Usage no npm install needed!

<script type="module">
  import mgfxConnectorRedis from 'https://cdn.skypack.dev/@mgfx/connector-redis';



mgFx: Managed Effects for JavaScript

mgFx is a library for JavaScript applications that allows you to define asynchronous, side-effecting code ('Tasks') and then invoke those tasks in a predictable and type-safe manner.

Quick-Start Guide

  1. Add mgFx to your existing NodeJS application:
$ yarn add mgfx
  1. Define and Implement a Task:
import { define, implement } from 'mgfx';

const greet = implement(
  define({ name: 'greet' }),
  (input) => `Hello ${input.name}!`
  1. Create a Connector:
import { localConnector } from 'mgfx';

const mgFx = localConnector();
  1. Tell your Connector to serve the Task implementation:

mgFx.run(greet({ name: 'World' })).value(console.log);

// Logs: `Hello World!`

Going Further

The Quick-Start guide merely scratches the surface of what mgFx can provide. Other features that need to be highlighted and documented are:

  • Run-time validation of Task input and output.
  • Composing Tasks that run other Tasks.
  • Run-time instrumentation and aggregated telemetry.
  • The additional tools in the mgFx ecosystem.


TODO: This section is currently out of date; it will be rewritten when telemetry aggregation has been *implemented.

A first-class priorty of mgFx is the ability to observe and analyse the behaviour of an application using mgFx at runtime. To achieve this, mgFx has three distinct 'layers' that are designed to work together:

  1. A built-in 'Instrumentation' system that exposes mgFx's runtime behaviour.
  2. @mgfx/recorder captures the Instrumenter's output and stores it in an SQL database.
  3. @mgfx/inspector provides a way for engineers to examine these events in a browser-based GUI.

This can also be depicted via a (somewhat crude) diagram:

+-------------------+        +----------------+           +--------------+        +-----------------+
| Your Application  | stdout |                |  INSERT/  |              | SELECT |                 |
|         +         +--------> @mgFx/recorder ------------> SQL Database ---------> @mgFx/inspector |
| mgFx Instrumenter |        |                |  UPDATE   |              |        |                 |
+-------------------+        +----------------+           +--------------+        +-----------------+

To enable Instrumentation for your mgFx application, simply create an Instrumenter instance, passing in your Scheduler instance. For example:

import { Inspector, singleProcess } from 'mgfx';

const { scheduler } = singleProcess({
  tasks: [
    /* Your list of tasks... */

new Instrumenter({ scheduler });

When you run your application, the Instrumenter will now emit structured messages to stdout. It is possible to pipe this output into @mgfx/recorder, like so:

$ node your-app.js | npx @mgfx/recorder sqlite mgfx-instrumentation.db

You may then use @mgfx/inspector to review events by pointing it to the same file:

$ npx @mgfx/inspector sqlite mgfx-instrumentation.db

Please consult the Inspector documentation for further information.