centrifuge
Advanced tools
JavaScript client SDK for bidirectional communication with Centrifugo and Centrifuge-based server from browser, NodeJS and React Native
This SDK provides a client to connect to Centrifugo or any Centrifuge-based server using pure WebSocket or one of the alternative transports (HTTP-streaming, SSE/EventSource, experimental WebTransport) from web browser, ReactNative, or NodeJS environments.
[!IMPORTANT]
This library behaves according to a common Centrifigo SDK spec. It's recommended to read that before starting to work with this SDK as the spec covers common SDK behavior - describes client and subscription state transitions, main options and methods. Then proceed with this readme for more specifics aboutcentrifuge-js.
The features implemented by this SDK can be found in SDK feature matrix.
centrifuge-jsv5.x is compatible with Centrifugo server v4 and v5 and Centrifuge >= 0.25.0. For Centrifugo v2, Centrifugo v3 and Centrifuge < 0.25.0 you should usecentrifuge-jsv2.x.
SDK can be installed via npm:
npm install centrifuge
And then in your project:
import { Centrifuge } from 'centrifuge';
In browser, you can import SDK from CDN (replace 5.0.0 with a concrete version number you want to use, see releases):
<script src="https://unpkg.com/centrifuge@5.0.0/dist/centrifuge.js"></script>
See also centrifuge-js on cdnjs. Note that centrifuge-js browser builds target ES6.
By default, library works with JSON only, if you want to send binary payloads go to Protobuf support section to see how to import client with Protobuf support.
The basic usage example may look like this:
// Use WebSocket transport endpoint.
const centrifuge = new Centrifuge('ws://centrifuge.example.com/connection/websocket');
// Allocate Subscription to a channel.
const sub = centrifuge.newSubscription('news');
// React on `news` channel real-time publications.
sub.on('publication', function(ctx) {
console.log(ctx.data);
});
// Trigger subscribe process.
sub.subscribe();
// Trigger actual connection establishement.
centrifuge.connect();
Note, that we explicitly call .connect() method to initiate connection establishement with a server and .subscribe() method to move Subscription to subsribing state (which should transform into subscribed state soon after connection with a server is established). The order of .connect() and .subscribe calls does not actually matter here.
Centrifuge object and Subscription object are both instances of EventEmitter. Below we will describe events that can be exposed in detail.
This SDK supports several real-time transports.
WebSocket is the main protocol used by centrifuge-js to communicate with a server.
In a browser environment WebSocket is available globally, but if you want to connect from NodeJS env – then you need to provide WebSocket constructor to centrifuge-js explicitly. See below more information about this.
It's the only transport for which you can just use a string endpoint as first argument of Centrifuge constructor. If you need to use other transports, or several transports – then you should use Array<TransportEndpoint>.
In the quick start example above we used WebSocket endpoint to configure Centrifuge. WebSocket is the main transport – it's bidirectional out of the box.
In some cases though, WebSocket connection may not be established (for example, due to corporate firewalls and proxies). For such situations centrifuge-js offers several WebSocket fallback options based on HTTP:
These two transports use Centrifugo/Centrifuge own bidirectional emulation layer. See more details in introduction post. Bidirectional emulation must be first enabled on a server-side. See Centrifugo docs to find out how.
After enabling HTTP-streaming or SSE endpoints on a server side you can slightly change client initialization and point Javascript SDK to a list of endpoints and transports you want to use:
const transports = [
{
transport: 'websocket',
endpoint: 'ws://example.com/connection/websocket'
},
{
transport: 'http_stream',
endpoint: 'http://example.com/connection/http_stream'
},
{
transport: 'sse',
endpoint: 'http://example.com/connection/sse'
}
];
const centrifuge = new Centrifuge(transports);
centrifuge.connect()
In this case, client will try transports in order, one by one, during the initial handshake. Until success. Then will only use a successfully chosen transport during reconnects.
Supported transports are:
websockethttp_streamssesockjs - SockJS can also be used as a fallback, SockJS is currently in DEPRECATED status in Centrifugal ecosystem. Also, sticky sessions must be used on the backend in distributed case with it. See more details belowwebtransport - this SDK also supports WebTransport in experimental form. See details belowIf you want to use sticky sessions on a load balancer level as an optimimization for Centrifugal bidirectional emulation layer keep in mind that we currently use same-origin credentials policy for emulation requests in http_stream and sse transport cases. So cookies will only be passed in same-origin case. Please open an issue in case you need to configure more relaxed credentials. Though in most cases stickyness based on client's IP may be sufficient enough.
SockJS usage is DEPRECATED in the Centrifugal ecosystem
If you want to use SockJS you must also import SockJS client before centrifuge.js
<script src="https://cdn.jsdelivr.net/npm/sockjs-client@1/dist/sockjs.min.js" type="text/javascript"></script>
<script src="https://unpkg.com/centrifuge@5.0.0/dist/centrifuge.js" type="text/javascript"></script>
Or provide it explicitly as a dependency:
import { Centrifuge } from 'centrifuge'
import SockJS from 'sockjs-client'
const transports = [{
transport: "sockjs",
endpoint: "http://localhost:8000/connection/sockjs"
}];
const centrifuge = new Centrifuge(transports, {
sockjs: SockJS
})
Note, that in SockJS case endpoint starts with http://, not with ws:// as we used above when connecting to a pure WebSocket endpoint.
WebTransport is experimental and is only supported by Centrifugo at the moment (i.e. it's not available in Centrifuge library for Go out of the box).
Server must be additionally configured to work with WebTransport connections – see information in Centrifugo WebTransport docs.
Let's look at top-level API of Centrifuge client.
As we already showed above, we must call connect() method to make an actual connection
request to Centrifugo server:
const centrifuge = new Centrifuge('ws://centrifuge.example.com/connection/websocket');
centrifuge.connect();
connect() triggers an actual connection request to server.
As soon as connection is established and client successfully authenticated – connected event on Centrifuge object instance will be called.
It's possible to listen to this event by setting event listener function on connected event:
centrifuge.on('connected', function(ctx) {
// now client connected to Centrifugo and authenticated.
});
connecting event fired when Centrifuge object goes to connecting state. This may be called during initial connect, or after being connected due to temporary connection loss.
centrifuge.on('connecting', function(ctx) {
// do whatever you need in case of connecting to a server
});
disconnected event fired on Centrifuge object every time client disconnects for some reason. This can be terminal disconnect due to advice from a server or disconnect initiated by client-side.
centrifuge.on('disconnected', function(ctx) {
// do whatever you need in case of disconnect from server
});
In some cases you may need to disconnect your client from server, use .disconnect() method to do this:
centrifuge.disconnect();
After calling this client will not try to reestablish connection periodically. You must call .connect() method manually again.
Sometimes you need to publish into channel without actually being subscribed to it. In this case you can use publish method:
centrifuge.publish("channel", {"input": "hello"}).then(function(res) {
console.log('successfully published');
}, function(err) {
console.log('publish error', err);
});
This is only valid for Centrifuge server library for Go and does not work for Centrifugo server at the moment. send method allows sending asynchronous message from a client to a server.
centrifuge.send({"input": "hello"}).then(function(res) {
console.log('successfully sent');
}, function(err) {
console.log('send error', err);
});
rpc method allows to send rpc request from client to server and wait for data response.
centrifuge.rpc("my.method.name", {"input": "hello"}).then(function(res) {
console.log('rpc result', res);
}, function(err) {
console.log('rpc error', err);
});
Allows to get history from a server. This is a top-level analogue of Subscription.history method. But accepts a channel as first argument.
centrifuge.history("channel", {since: {offset: 0, epoch: "xyz"}, limit: 10}).then(function(resp) {
console.log(resp);
}, function(err) {
console.log('history error', err);
});
Allows to get presence info from a server. This is a top-level analogue of Subscription.presence method. But accepts a channel as first argument.
centrifuge.presence("channel").then(function(resp) {
console.log(resp);
}, function(err) {
console.log('presence error', err);
});
Allows to get presence stats from a server. This is a top-level analogue of Subscription.presenceStats method. But accepts a channel as first argument.
centrifuge.presenceStats("channel").then(function(resp) {
console.log(resp);
}, function(err) {
console.log('presence stats error', err);
});
Returns a Promise which will be resolved upon connection establishement (i.e. when Client goes to connected state).
setToken may be useful to dynamically change the connection token. For example when you need to implement login/logout workflow. See an example in blog post.
To listen asynchronous error happening internally while Centrifuge client works you can set an error handler:
const centrifuge = new Centrifuge('ws://centrifuge.example.com/connection/websocket');
centrifuge.on('error', function(ctx) {
console.log(ctx);
});
This can help you to log failed connection attempts, or token refresh errors, etc.
Depending on authentication scheme used by a server you may also want to provide connection token:
const centrifuge = new Centrifuge('ws://centrifuge.example.com/connection/websocket', {
token: '<CONNECTION_TOKEN>'
});
In case of Centrifugo on a server side this may be a JSON Web Token - see authentication documentation for details on how to generate it on your backend side.
Connection token must come to the frontend from application backend - i.e. must be generated on the backend side. The way to deliver token to the application frontend is up to the developer. Usually you can pass it in template rendering context or issue a separate call to request a connection token from the backend.
If the token sets connection expiration then the client SDK will keep the token refreshed. It does this by calling a special callback function. This callback must return a new token. If a new token with updated connection expiration is returned from callback then it's sent to Centrifugo. In case of error returned by your callback SDK will retry the operation after some jittered time. You can throw a special error (throw new Centrifuge.UnauthorizedError();) from getToken function to move the client into disconnected state (for example, when there is no permission to connect anymore).
An example of possible getToken function implementation:
import { Centrifuge, UnauthorizedError } from 'centrifuge';
async function getToken() {
if (!loggedIn) {
return "";
}
const res = await fetch('/centrifuge/connection_token');
if (!res.ok) {
if (res.status === 403) {
// Return special error to not proceed with token refreshes, client will be disconnected.
throw new UnauthorizedError();
}
// Any other error thrown will result into token refresh re-attempts.
throw new Error(`Unexpected status code ${res.status}`);
}
const data = await res.json();
return data.token;
}
const client = new Centrifuge(
'ws://localhost:8000/connection/websocket',
{
token: 'JWT-GENERATED-ON-BACKEND-SIDE',
getToken: getToken
}
);
If initial token is not provided, but
getTokenis specified – then SDK assumes that developer wants to use token authentication. In this case SDK attempts to get a connection token before establishing an initial connection.
What we usually want from Centrifugo is to receive new messages published into channels. To do this we must create Subscription object.
The simplest usage that allow to subscribe on channel and listen to new messages is:
const sub = centrifuge.newSubscription('example');
sub.on('publication', function(ctx) {
// handle new Publication data coming from channel "news".
console.log(ctx.data);
});
sub.subscribe();
Some events which can be listened on Subscription object are:
publication – called when new publication received from a Subscription channeljoin – called when someone joined channelleave – called when someone left channelsubscribing - called when Subscription goes to subscribing state (initial subscribe and re-subscribes)subscribed – called when Subscription goes to subscribed stateunsubscribed – called when Subscription goes to unsubscribed stateerror – called when subscription on channel failed with error. It can be called several times
during lifetime as browser client automatically resubscribes on channels after successful reconnect
(caused by temporary network disconnect for example or Centrifugo server restart)Don't be frightened by amount of events available. In most cases you only need some of them until you need full control to what happens with your subscriptions.
Subscription objects are instances of EventEmitter.
presence allows to get information about clients which are subscribed on channel at
this moment. Note that this information is only available if presence option enabled
in Centrifugo configuration for all channels or for channel namespace.
const sub = centrifuge.newSubscription("news");
sub.subscribe()
sub.presence().then(function(ctx) {
console.log(ctx.clients);
}, function(err) {
// presence call failed with error
});
presence is internally a promise that will be waiting for subscription subscribe success if required.
As you can see presence data is a map where keys are client IDs and values are objects with client information.
Format of err in error callback:
{
"code": 108,
"message": "not available"
}
code - error code (number)message – error description (string)Note, that in order presence to work corresponding options must be enabled in server channel configuration (on top level or for channel namespace)
presenceStats allows to get two counters from a server: number of total clients currently subscribed and number of unique users currently subscribed. Note that this information is only available if presence option enabled in server configuration for a channel.
sub.presenceStats().then(function(ctx) {
console.log(ctx.numClients);
}, function(err) {
// presence stats call failed with error
});
history method allows to get last messages published into channel. Note that history for channel must be configured in Centrifugo to be available for history calls from client.
sub.history({limit: 100}).then(function(ctx) {
console.log(ctx.publications);
}, function(err) {
// history call failed with error
});
Note, that in order history to work corresponding options must be enabled in server channel configuration (on top level or for channel namespace)
Some history options available:
limit (number)since (StreamPosition)reverse (boolean)resp = await subscription.history({'since': {'offset': 2, 'epoch': 'xcf4w'}, limit: 100});
If server can't fulfill a query for history (due to stream retention - size or expiration, or malformed offset, or stream already has another epoch) then an Unrecoverable Position Error will be returned (code 112).
To only call for current offset and epoch use:
resp = await subscription.history({limit: 0});
I.e. not providing since and using zero limit.
publish method of Subscription object allows publishing data into channel directly from a client.
Using client-side publish is not an idiomatic Centrifugo usage in many cases. Centrifugo is standalone server and when publishing from a client you won't get the message on the backend side (except using publish proxy feature of Centrifugo). In most real-life apps you need to send new data to your application backend first (using the convenient way, for example AJAX request in web app) and then publish data to Centrifugo over Centrifugo API.
Just like presence and history publish must be allowed in Centrifugo configuration for all channels or for channel namespace.
sub.publish({"input": "hello world"}).then(function() {
// success ack from Centrifugo received
}, function(err) {
// publish call failed with error
});
});
Note, that in order publish to work in Centrifugo corresponding option must be enabled in server channel configuration or client should have capability to publish.
You can call unsubscribe method to unsubscribe from a channel:
sub.unsubscribe();
Important thing to mention is that unsubscribing from subscription does not remove event handlers you already set to that Subscription object. This allows to simply subscribe to channel again later calling .subscribe() method of subscription (see below). But there are cases when your code structured in a way that you need to remove event handlers after unsubscribe to prevent them be executed twice in the future. To do this remove event listeners explicitly after calling unsubscribe():
sub.unsubscribe();
sub.removeAllListeners();
Returns a Promise which will be resolved upon subscription success (i.e. when Subscription goes to subscribed state).
You may want to provide subscription token:
const sub = centrifuge.newSubscription("news", {
token: '<SUBSCRIPTION_TOKEN>'
});
In case of Centrifugo on a server side this may be a JSON Web Token - see channel token auth documentation for details on how to generate it on your backend side.
Subscription token must come to the frontend from application backend - i.e. must be generated on the backend side. The way to deliver token to the application frontend is up to the developer. Usually you can pass it in template rendering context or issue a separate call to request a connection token from the backend.
If token sets subscription expiration client SDK will keep token refreshed. It does this by calling special callback function. This callback must return a new token. If new token with updated subscription expiration returned from a calbback then it's sent to Centrifugo. If your callback returns an empty string – this means user has no permission to subscribe to a channel anymore and subscription will be unsubscribed. In case of error returned by your callback SDK will retry operation after some jittered time.
An example:
import { Centrifuge, UnauthorizedError } from 'centrifuge';
async function getToken(ctx) {
// ctx argument has a channel.
const res = await fetch('/centrifuge/subscription_token', {
method: 'POST',
headers: new Headers({ 'Content-Type': 'application/json' }),
body: JSON.stringify(ctx)
});
if (!res.ok) {
if (res.status === 403) {
// Return special error to not proceed with token refreshes, subscription will be unsubscribed.
throw new UnauthorizedError();
}
// Any other error thrown will result into token refresh re-attempts.
throw new Error(`Unexpected status code ${res.status}`);
}
const data = await res.json();
return data.token;
}
const client = new Centrifuge('ws://localhost:8000/connection/websocket', {});
const sub = centrifuge.newSubscription(channel, {
token: 'JWT-GENERATED-ON-BACKEND-SIDE',
getToken: getToken,
});
sub.subscribe();
If initial token is not provided, but
getTokenis specified – then SDK assumes that developer wants to use token authorization for a channel subscription. In this case SDK attempts to get a subscription token before initial subscribe.
According to client SDK spec centrifuge-js supports several methods to manage client-side subscriptions in internal registry. The following methods are available on top level of the Centrifuge SDK client instance.
newSubscription(channel: string, options?: Partial<SubscriptionOptions>): Subscription
Creates new Subscription to a channel or throws an exception if the Subscription to a channel already exists in the internal registry of the client.
getSubscription(channel: string): Subscription | null
getSubscription returns Subscription if it's registered in the internal registry or null.
removeSubscription(sub: Subscription | null)
removeSubscription allows removing Subcription from the internal registry. Subscrption must be in unsubscribed state.
subscriptions(): Record<string, Subscription>
Get a map with all current client-side subscriptions registered in the client.
There is also a command batching support. It allows to send several commands to a server in one request - may be especially useful when connection established via one of HTTP-based transports.
You can start collecting commands by calling startBatching() method:
centrifuge.startBatching();
Finally if you don't want batching anymore call stopBatching() method:
centrifuge.stopBatching();
This call will flush all collected commands to a network.
We encourage using client-side subscriptions where possible as they provide a better control and isolation from connection. But in some cases you may want to use server-side subscriptions (i.e. subscriptions created by server upon connection establishment).
Technically, client SDK keeps server-side subscriptions in internal registry (similar to client-side subscriptions but without possibility to control them).
To listen for server-side subscription events use callbacks as shown in example below:
const client = new Centrifuge('ws://localhost:8000/connection/websocket', {});
client.on('subscribed', function(ctx) {
// Called when subscribed to a server-side channel upon Client moving to
// connected state or during connection lifetime if server sends Subscribe
// push message.
console.log('subscribed to server-side channel', ctx.channel);
});
client.on('subscribing', function(ctx) {
// Called when existing connection lost (Client reconnects) or Client
// explicitly disconnected. Client continue keeping server-side subscription
// registry with stream position information where applicable.
console.log('subscribing to server-side channel', ctx.channel);
});
client.on('unsubscribed', function(ctx) {
// Called when server sent unsubscribe push or server-side subscription
// previously existed in SDK registry disappeared upon Client reconnect.
console.log('unsubscribed from server-side channel', ctx.channel);
});
client.on('publication', function(ctx) {
// Called when server sends Publication over server-side subscription.
console.log('publication receive from server-side channel', ctx.channel, ctx.data);
});
client.connect();
Server-side subscription events mostly mimic events of client-side subscriptions. But again – they do not provide control to the client and managed entirely by a server side.
Additionally, Client has several top-level methods to call with server-side subscription related operations:
publish(channel, data)history(channel, options)presence(channel)presenceStats(channel)You can check out all available options with description in source code.
Let's look at available configuration parameters when initializing Centrifuge object instance.
debug is a boolean option which is false by default. When enabled lots of various debug
messages will be logged into javascript console. Mostly useful for development or
troubleshooting.
When client disconnected from a server it will automatically try to reconnect using a backoff algorithm with jitter. minReconnectDelay option sets minimal interval value in milliseconds before first reconnect attempt. Default is 500 milliseconds.
maxReconnectDelay sets an upper reconnect delay value. Default is 20000 milliseconds - i.e. clients won't have delays between reconnect attempts which are larger than 20 seconds.
maxServerPingDelay sets the maximum delay of server pings after which connection is considered broken and client reconnects. In milliseconds. Default is 10000.
Set initial connection token.
Set function for getting connection token. This may be used for initial token loading and token refresh mechanism (when initial token is going to expire).
Set custom data to send to a server withing every connect command.
Set custom client name. By default, it's set to js. This is useful for analitycs and semantically must identify an environment from which client establishes a connection.
Version of your application - useful for analitycs.
Timeout for operations in milliseconds.
websocket option allows to explicitly provide custom WebSocket client to use. By default centrifuge-js will try to use global WebSocket object, so if you are in web browser – it will just use native WebSocket implementation. See notes about using centrifuge-js with NodeJS below.
sockjs option allows to explicitly provide SockJS client object to Centrifuge client.
To import client which uses Protobuf protocol under the hood:
<script src="https://unpkg.com/centrifuge@5.0.0/dist/centrifuge.protobuf.js"></script>
Or if you are developing with npm:
import { Centrifuge } from 'centrifuge/build/protobuf';
This client uses protobuf.js under the hood.
When running with Protobuf-based client, you can send and receive any binary data as Uint8Array. Make sure data is properly encoded when calling methods of Centrifuge Protobuf-based instance. For example, you can not just send JSON-like objects like in JSON protocol case, you need to encode data to Uint8Array first:
const data = new TextEncoder("utf-8").encode(JSON.stringify({"any": "data"}));
sub.publish(data);
NodeJS does not have native WebSocket library in std lib. To use centrifuge-js on Node you need to explicitly provide WebSocket constructor to the library.
First, install WebSocket dependency:
npm install ws
At this point you have 2 options. Explicitly pass WebSocket object to Centrifuge.
import { Centrifuge } from 'centrifuge';
import WebSocket from 'ws';
var centrifuge = new Centrifuge('ws://localhost:8000/connection/websocket', {
websocket: WebSocket
})
Or define it globally:
import { Centrifuge } from 'centrifuge';
import WebSocket from 'ws';
global.WebSocket = WebSocket;
const centrifuge = new Centrifuge('ws://localhost:8000/connection/websocket');
If you are building a client for a non-browser environment and want to pass custom headers then you can use the following approach to wrap a WebSocket constructor and let custom options to be used on connection initialization:
const myWs = function (options) {
return class wsClass extends WebSocket {
constructor(...args) {
if (args.length === 1) {
super(...[...args, 'centrifuge-json', ...[options]])
} else {
super(...[...args, ...[options]])
}
}
}
}
It should be now possible to use pass your custom WebSocket constructor to centrifuge-js and so custom headers will be used when connecting to a server (only in non-browser environment):
var centrifuge = new Centrifuge('ws://localhost:8000/connection/websocket', {
websocket: myWs({ headers: { Authorization: '<token or key>' } }),
});
See a basic example with React Native where this technique is used in this comment.
If you want to run centrifuge-js tests locally, start test Centrifugo server:
docker compose up
Then:
yarn test
FAQs
JavaScript client SDK for bidirectional communication with Centrifugo and Centrifuge-based server from browser, NodeJS and React Native
The npm package centrifuge receives a total of 45,210 weekly downloads. As such, centrifuge popularity was classified as popular.
We found that centrifuge demonstrated a healthy version release cadence and project activity because the last version was released less than a year ago. It has 0 open source maintainers collaborating on the project.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Security News
In this segment of the Risky Business podcast, Feross Aboukhadijeh and Patrick Gray discuss the challenges of tracking malware discovered in open source softare.
Research
Security News
A threat actor's playbook for exploiting the npm ecosystem was exposed on the dark web, detailing how to build a blockchain-powered botnet.
Security News
NVD’s backlog surpasses 20,000 CVEs as analysis slows and NIST announces new system updates to address ongoing delays.