https://github.com/micro-tools/node-pohl

distributed redis queue messaging, makes sure that only a single instance actually works on the task, sequential callback implementation
https://github.com/micro-tools/node-pohl

callback-style fast messages redis sequential

Last synced: 8 months ago
JSON representation

distributed redis queue messaging, makes sure that only a single instance actually works on the task, sequential callback implementation

• Host: GitHub
• URL: https://github.com/micro-tools/node-pohl
• Owner: micro-tools
• License: mit
• Created: 2016-08-30T15:10:45.000Z (over 9 years ago)
• Default Branch: master
• Last Pushed: 2022-04-14T05:07:34.000Z (over 3 years ago)
• Last Synced: 2025-04-01T07:24:21.552Z (9 months ago)
• Topics: callback-style, fast, messages, redis, sequential
• Language: JavaScript
• Homepage:
• Size: 198 KB
• Stars: 4
• Watchers: 1
• Forks: 2
• Open Issues: 6
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# node-pohl

[![Build Status](https://travis-ci.org/krystianity/node-pohl.svg?branch=master)](https://travis-ci.org/krystianity/node-pohl)

- reliable distributed rpc messages via redis in simple callback-style

- ES6, coverage 66%+, lightweight, scalable & fast

- uses pub & sub, but makes sure that only a single instance actually works on the task via redlock algorithm.

- kind of rpc library that feels like making simple callbacks except for the fact that you can make them between services and not classes

- simplistic circuit breaker to prevent timeout waves

- metric events

- pauseable receiver (.pause(), .resume()) for green/blue deployment scenarios

#use case

- image you have enterprise microservices that have to make rpc/rest/soap calls to dispatch other information during incoming requests

- now think of the potential overhead that protocols like http can cause, you might most likely have to wait for 20-150ms overhead might depend on your infrastructure

- there has to be a way to have a standing connection, that sends messages reliably, distributed, scalable and fast with a fixed overhead that can be calculated

- all you need is a redis cluster/sentinel setup (works with single instance as well..), include this lib in your services and call 2 functions thats about it

- message queuing, task locking, failovers or timeouts have been wrapped in a super simple callback-like syntax that also scales inside of your software to multiple topics and endless message/task types

- overhead is between 3-5ms constantly

- benchmarks hits 7500 rpc/s (full-roundtrip) on mobile i7 @ 2GHZ and Redis 3.2.1 single docker instance, with a single sender

#how to use/install

- `npm install pohl`

- check ./example/index.js for a usage example

- run example with `npm start`

- run tests with `npm test` (requires localhost redis with default conf)

- run benchmark with `npm run benchmark`

#other

- License: MIT

- Author: Christian Fröhlingsdorf