A WebSocket JavaScript library.
npm i @anephenix/sarus
After installing Sarus, you can use the client library with your frontend codebase:
import Sarus from '@anephenix/sarus';
const sarus = new Sarus({
url: 'wss://ws.anephenix.com',
});
Sarus creates a WebSocket connection to the url. You can then attach event
listener functions to that WebSocket client via sarus
for events like:
Here's an example of attaching events on client initialization:
// Log a message that the connection is open
const noteOpened = () => console.log('Connection opened');
// Assuming that the WebSocket server is sending JSON data,
// you can use this to parse the data payload;
const parseMessage = (event) => {
const message = JSON.parse(event.data);
// Then do what you like with the message
};
// Log a message that the connection has closed
const noteClosed = () => console.log('Connection closed');
// If an error occurs, throw the error
const throwError = (error) => throw error;
// Create the Sarus instance with the event listeners
const sarus = new Sarus({
url: 'wss://ws.anephenix.com',
eventListeners: {
open: [noteOpened],
message: [parseMessage],
close: [noteClosed],
error: [throwError],
},
});
You can specify all of the event listeners at initialisation, or just one of them:
// Assuming that the WebSocket server is sending JSON data,
// you can use this to parse the data payload;
const parseMessage = (event) => {
const message = JSON.parse(event.data);
// Then do what you like with the message
};
// Create the Sarus instance with the event listeners
const sarus = new Sarus({
url: 'wss://ws.anephenix.com',
eventListeners: {
message: [parseMessage],
},
});
You can also add eventListeners after client initialization:
/*
A function that stores messages in the browser's LocalStorage, possibly
for debugging, or for event stream processing on the client side.
*/
const storeMessage = (event) => {
const store = window.localStorage;
let record = store.getItem('messages');
if (!record) {
record = [];
} else {
record = JSON.parse(record);
}
record.push(event.data);
store.setItem('messages', JSON.stringify(record));
};
// Attach the storeMessage function to Sarus when it receives a message from
// the WebSocket server
sarus.on('message', storeMessage);
You can also use it to send messages to the WebSocket server:
sarus.send('Hello world');
WebSockets can close unexpectedly. When a WebSocket instance is closed, it
cannot be reopened. To re-establish a WebSocket connection, you have to create
a new WebSocket
instance to replace the closed instance.
Usually you would handle this by writing some JavaScript to wrap the WebSocket interface, and trigger opening a new WebSocket connection upon a close event occurring.
Sarus will do this automatically for you.
It does this by attaching a connect
function on the close
event happening
on the WebSocket
instance. If the WebSocket
instance closes, the connect
function will simply create a new WebSocket
instance with the same
parameters that were passed to the previous instance.
The connect
function is called immediately by default, and it will repeat
this until it gets a WebSocket
instance whose connection is open.
If you do not want the WebSocket to reconnect automatically, you can pass the
reconnectAutomatically
parameter into the sarus client at the point of
initializing the client, like the example below.
const sarus = new Sarus({
url: 'wss://ws.anephenix.com',
reconnectAutomatically: false,
});
There may be a case where you wish to close a WebSocket connection (such as when logging out of a service). Sarus provides a way to do that:
sarus.disconnect();
Calling that function on the sarus client will do 2 things:
reconnectAutomatically
flag to false.Event listeners listening on the WebSocket's close event will still trigger, but the client will not attempt to reconnect automatically.
If you wish to close the WebSocket but not override the reconnectAutomatically
flag, pass this:
sarus.disconnect(true);
The client will attempt to reconnect automatically.
When a connection is severed and the sarus client tries to reconnect automatically, it will do so with a delay of 1000ms (1 second).
If you pass a number, then it will delay the reconnection attempt by that time (in miliseconds):
const sarus = new Sarus({
url: 'wss://ws.anephenix.com',
retryConnectionDelay: 500, // equivalent to 500ms or 1/2 second
});
NOTE
In the past this option needed to be explicitly passed, but we decided to change it to be enabled by default. Without it, any disconnection could result in thousands of attempted reconnections by one client in the space of a few seconds.
When a WebSocket connection is closed, any functions attached to events emitted by that WebSocket instance need to be attached to the new WebSocket instance. This means that you end up writing some JavaScript that handles attaching event listener functions to new WebSocket instances when they get created to replace closed instances.
Sarus does this for you. You have 2 ways to attach functions to your WebSocket event listeners - either when creating the Sarus instance, or after it exists:
// Log a message that the connection is open
const noteOpened = () => console.log('Connection opened');
// Assuming that the WebSocket server is sending JSON data,
// you can use this to parse the data payload;
const parseMessage = (event) => {
const message = JSON.parse(event.data);
// Then do what you like with the message
};
// Log a message that the connection has closed
const noteClosed = () => console.log('Connection closed');
// If an error occurs, throw the error
const throwError = (error) => throw error;
// Create the Sarus instance with the event listeners
const sarus = new Sarus({
url: 'wss://ws.anephenix.com',
eventListeners: {
open: [noteOpened],
message: [parseMessage],
close: [notedClosed],
error: [throwError],
},
});
In this example, those functions will be bound to the WebSocket instance. If the WebSocket instance's connection closes, a new WebSocket instance is created by Sarus to reconnect automatically. The event listeners set in Sarus will be attached to that new WebSocket instance automatically.
That is one way that Sarus allows you to bind event listeners to events on the
WebSocket connection. Another way to do it is to call the on
function on the
Sarus instance, like this:
/*
A function that stores messages in the browser's LocalStorage, possibly
for debugging, or for event stream processing on the client side.
*/
const storeMessage = (event) => {
const store = window.localStorage;
let record = store.getItem('messages');
if (!record) {
record = [];
} else {
record = JSON.parse(record);
}
record.push(event.data);
store.setItem('messages', JSON.stringify(record));
};
// Attach the storeMessage function to Sarus when it receives a message from
// the WebSocket server
sarus.on('message', storeMessage);
If you want to remove a function from a WebSocket event listener, you can do
that by calling the off
function on Sarus like this:
// Pass the function variable
sarus.off('message', storeMessage);
// You can also pass the name of the function as well
sarus.off('message', 'storeMessage');
If you attempt to remove an event listener function which is not in the list of event listeners, then an error will be thrown by Sarus. This is a deliberate behaviour of Sarus. Rather than silently failing to remove a function because it was not there (or perhaps there was a misspelling of the function name), it will explicitly throw an error, so that the developer can be made aware of it and handle it as they wish.
If the developer is happy for an event listener removal to fail without
throwing an error, they can pass this to the off
function:
sarus.off('message', 'myNonExistentFunction', { doNotThrowError: true });
Sending a message from a Websocket client to the server depends on the WebSocket connection being open. If the connection is closed, then you will need to either prevent the messages from being sent (block message delivery), or you will need to queue the messages for delivery (queue message delivery). Either option requires writing some JavaScript to do that.
To handle this case, Sarus implements a client-based message queue, so that messages are sent only when there is an open WebSocket connection.
The message queue is stored in memory. If the web page is refreshed, then the messages in the queue will be lost. If you want to persist the messages in the queue between web page refreshes, you can pass an option to Sarus to have the messages stored using the sessionStorage protocol:
const sarus = new Sarus({
url: 'wss://ws.anephenix.com',
storageType: 'session',
});
The sessionStorage protocol guarantees that messages are stored between web page refreshes, but only in the context of that web page's browser tab or window. The messages will not persist in new browser tabs/windows, or after the browser has been closed.
If you want the storage of those messages to persist beyond web page sessions, then you can use the localStorage protocol as the storage mechanism:
const sarus = new Sarus({
url: 'wss://ws.anephenix.com',
storageType: 'local',
});
LocalStorage guarantees that the messages are persisted beyond browsers being closed and reopened, as well as when the page is opened in a new tab/window.
NOTE When persisting messages, be careful that the messages are safe to persist in browser storage, and do not contain sensitive information. If you want messages to be wiped when the user closes the browser, use 'session' as the storage type.
NOTE Each web browser implements arbitrary limits for how much data can be stored in sessionStorage/localStorage for a domain. When that limit is reached, the web browser will throw a QUOTA_EXCEEDED_ERR error. The limits tend to be in the 5MB-10MB range, but do vary between browsers.
If you think that there is a potential case for you ending up queuing at least
5MB of data in messages to send to a WebSocket server, then you may want to
wrap sarus.send
function calls in a try/catch statement, so as to handle
those messages, should they occur.
Configure exponential backoff like so:
import Sarus from '@anephenix/sarus';
const sarus = new Sarus({
url: 'wss://ws.anephenix.com',
exponentialBackoff: {
// Exponential factor, here 2 will result in
// 1 s, 2 s, 4 s, and so on increasing delays
backoffRate: 2,
// Never wait more than 2000 seconds
backoffLimit: 2000,
},
});
When a connection attempt repeatedly fails, decreasing the delay exponentially between each subsequent reconnection attempt is called Exponential backoff. The idea is that if a connection attempt failed after 1 second, and 2 seconds, then it is not necessary to check it on the 3rd second, since the probability of a reconnection succeeding on the third attempt is most likely not going up. Therefore, increasing the delay between each attempt factors in the assumption that a connection is not more likely to succeed by repeatedly probing in regular intervals.
This decreases both the load on the client, as well as on the server. For a client, fewer websocket connection attempts decrease the load on the client and on the network connection. For the server, should websocket requests fail within, then the load for handling repeatedly failing requests will fall as well. Furthermore, the burden on the network will also be decreased. Should for example a server refuse to accept websocket connections for one client, then there is the possibility that other clients will also not be able to connect.
Sarus implements truncated exponential backoff, meaning that the maximum
reconnection delay is capped by another factor backoffLimit
and will never
exceed it. The exponential backoff rate itself is determined by backoffRate
.
If backoffRate
is 2, then the delays will be 1 s, 2 s, 4 s, and so on.
The algorithm for reconnection looks like this in pseudocode:
// Configurable
const backoffRate = 2;
// The maximum delay will be 400s
const backoffLimit = 400;
let notConnected = false;
let connectionAttempts = 1;
while (notConnected) {
const delay = Math.min(
Math.pow(connectionAttempts, backoffRate),
backoffLimit,
);
await delay(delay);
notConnected = tryToConnect();
connectionAttempts += 1;
}
Sarus has a number of other options that you can pass to the client during initialization. They are listed in the example below:
const sarus = new Sarus({
url: 'wss.anephenix.com',
protocols: 'hybi-00',
retryProcessTimePeriod: 25,
storageKey: 'messageQueue',
});
The protocols
property is used to specify the sub-protocol that the WebSocket
connection should use. You can pass either a string, or an array of strings.
The retryProcessTimePeriod
property is used to help buffer the time between
trying to resend a message over a WebSocket connection. By default it is a
number, 50 (for 50 miliseconds). You can adjust this value in the client
instance.
The storageKey
property is a key that is used with sessionStorage and
localStorage to store and retrieve the messages in the message queue. By
default it is set to 'sarus'. You can set this to another string value if
you wish. You can also inspect the message queue independently of Sarus by
making calls to the sessionStorage/localStorage api with that key.
The code for the library is written using ES2015 features, and the idea is that developers can directly load that code into their application, rather than loading it as an external dependency in a transpiled and minified format.
This gives the developer the freedom to use it as they wish with the frontend tools that they use, be it Babel, WebPack, Rollup, or even Browserify.
npm t
This will run tests using jest and with code coverage enabled.
© 2020 Anephenix OÜ. Sarus is licensed under the MIT License.