Seneca-Web-Helper

Two useful plugins for using seneca-web in a distributed manner.

This uses the magic of redis-pubsub and fire/forget mechanisms to mount routes when services come online.

The goal here is to stop excessive routes from being mounted to the target host server, and to ensure resilisiancy when new hosts or clients come online.

This plugins acts as a proxy to seneca-web calls, and uses redis pubsub to ensure all hosts receive all routes only when they need to.

Seneca compatibility

Supports Seneca versions 1.x - 3.x

Important Notes

• the seneca-redis-transport transport is currently broken. Use seneca-redis-transport-fork

• the cache mechanism is based upon the tag of each client microservice - this must be set.

• hostname defaults to the OS's hostname and is used to ensure targetted population of routes for each host under a given hostname - if this is the same for multiple hosts, you'll end up with a lot of noise.

• if a microservice comes online with new routes and all hosts are already aware of it, the new routes will not be mounted. You'll need to call into role:web,clear:all to get all hosts to clear their cache and repopulate routes.

Install

To install, simply use npm. You'll need to install seneca, seneca-web, and seneca-redis-transport-fork if you haven't already.

npm i seneca seneca-web seneca-redis-transport-fork

Host

const os = require('os')
const Express = require('express')
const Seneca = require('seneca')
const SenecaWeb = require('seneca-web')
const SenecaWebAdapterExpress = require('seneca-web-adapter-express')
const {WebHost} = require('seneca-web-helper')
const SenecaRedisTransport = require('seneca-redis-transport-fork')

const app = Express()
const server = http.createServer(app)

Seneca({ tag: 'api-gateway' })

  .use(SenecaRedisTransport)

  .use(SenecaWeb, {
    adapter: SenecaWebAdapterExpress,
    options: {parseBody: false},
    context: new Express.Router(),
    routes: []
  }))

  .use(WebHost)

  // you need to provide some kind of point->point mechanism for web requests
  .client({type: 'tcp', pins: ['role:test,cmd:*']})

  .ready(function () {

      this.act('role:web,register:all', {to_hostname: os.hostname()}) // this is the magic

      app.use('/api', this.export('web/context')())

      server.listen(3000, () => this.log.info('api gateway listening!'))

  })

Client

const Seneca = require('seneca')
const SenecaWeb = require('seneca-web')
const {WebClient} = require('seneca-web-helper')
const SenecaRedisTransport = require('seneca-redis-transport-fork')

Seneca({tag: 'microservice-1'})

  .use(SenecaRedisTransport)

  .use(WebClient, {
    routes: [{
      pin: 'role:test,cmd:*',
      map: {test: true}
    }]
  })

  // you need to provide some kind of point->point mechanism for web requests
  .listen({type: 'tcp', pins: ['role:test,cmd:*']})
  .add('role:test,cmd:foo', (args, cb) => cb(null, {ok: true}))

  .ready(function () {
    this.act('role:web', {register: this.options().tag}) // this is the magic
    this.log.info('microservice ready!')
  })

Contributing

The Senecajs org encourages open participation. If you feel you can help in any way, be it with documentation, examples, extra testing, or new features please get in touch.

License

Copyright Tyler Waters and other contributors 2019, Licensed under MIT.