A Node.js client for the NATS messaging system.
npm install nats
var NATS = require('nats');
var nats = NATS.connect();
// Simple Publisher
nats.publish('foo', 'Hello World!');
// Simple Subscriber
nats.subscribe('foo', function(msg) {
console.log('Received a message: ' + msg);
});
// Unsubscribing
var sid = nats.subscribe('foo', function(msg) {});
nats.unsubscribe(sid);
// Subscription/Request callbacks are given multiple arguments:
// - msg is the payload for the message
// - reply is an optional reply subject set by the sender (could be undefined)
// - subject is the subject the message was sent (which may be more specific
// than the subscription subject - see "Wildcard Subscriptions".
// - finally the subscription id is the local id for the subscription
// this is the same value returned by the subscribe call.
nats.subscribe('foo', function(msg, reply, subject, sid) {
if(reply) {
nats.publish(reply, "got " + msg + " on " + subject + " in subscription id " + sid);
return;
}
console.log('Received a message: ' + msg + " it wasn't a request.");
});
// Request Streams
var sid = nats.request('request', function(response) {
console.log('Got a response in msg stream: ' + response);
});
// Request with Auto-Unsubscribe. Will unsubscribe after
// the first response is received via {'max':1}
nats.request('help', null, {'max':1}, function(response) {
console.log('Got a response for help: ' + response);
});
// Request for single response with timeout.
nats.requestOne('help', null, {}, 1000, function(response) {
// `NATS` is the library.
if(response instanceof NATS.NatsError && response.code === NATS.REQ_TIMEOUT) {
console.log('Request for help timed out.');
return;
}
console.log('Got a response for help: ' + response);
});
// Replies
nats.subscribe('help', function(request, replyTo) {
nats.publish(replyTo, 'I can help!');
});
// Close connection
nats.close();
The json
connect property makes it easier to exchange JSON data with other
clients.
var nc = NATS.connect({json: true});
nc.on('connect', function() {
nc.on('error', function(err) {
console.log(err);
});
nc.subscribe("greeting", function(msg, reply) {
// msg is a parsed JSON object object
if(msg.name && msg.reply) {
nc.publish(reply, {greeting: "hello " + msg.name});
}
});
// As with all inputs from unknown sources, if you don't trust the data
// you should verify it prior to accessing it. While JSON is safe because
// it doesn't export functions, it is still possible for a client to
// cause issues to a downstream consumer that is not written carefully
nc.subscribe("unsafe", function(msg) {
// for example a client could inject a bogus `toString` property
// which could cause your client to crash should you try to
// concatenation with the `+` like this:
// console.log("received", msg + "here");
// `TypeError: Cannot convert object to primitive value`
// Note that simple `console.log(msg)` is fine.
if (msg.hasOwnProperty('toString')) {
console.log('tricky - trying to crash me:', msg.toString);
return;
}
// of course this is no different than using a value that is
// expected in one format (say a number), but the client provides
// a string:
if (isNaN(msg.amount) === false) {
// do something with the number
}
//...
});
// the bad guy
nc.publish("unsafe", {toString: "no good"});
nc.flush(function() {
nc.close();
});
});
// "*" matches any token, at any level of the subject.
nats.subscribe('foo.*.baz', function(msg, reply, subject) {
console.log('Msg received on [' + subject + '] : ' + msg);
});
nats.subscribe('foo.bar.*', function(msg, reply, subject) {
console.log('Msg received on [' + subject + '] : ' + msg);
});
// ">" matches any length of the tail of a subject, and can only be
// the last token E.g. 'foo.>' will match 'foo.bar', 'foo.bar.baz',
// 'foo.foo.bar.bax.22'
nats.subscribe('foo.>', function(msg, reply, subject) {
console.log('Msg received on [' + subject + '] : ' + msg);
});
// All subscriptions with the same queue name will form a queue group.
// Each message will be delivered to only one subscriber per queue group,
// queuing semantics. You can have as many queue groups as you wish.
// Normal subscribers will continue to work as expected.
nats.subscribe('foo', {'queue':'job.workers'}, function() {
received += 1;
});
var nats = require('nats');
var servers = ['nats://nats.io:4222', 'nats://nats.io:5222', 'nats://nats.io:6222'];
// Randomly connect to a server in the cluster group.
var nc = nats.connect({'servers': servers});
// currentServer is the URL of the connected server.
console.log("Connected to " + nc.currentServer.url.host);
// Preserve order when connecting to servers.
nc = nats.connect({'dontRandomize': true, 'servers':servers});
// Unsubscribing removes the subscription handler for a subscription
// and cancels the subscription. Any pending messages on the client's
// buffer are discarded.
//
// Draining is similar to unsubscribe, but the client instead
// sends the unsubscribe request followed by a flush. When the flush
// returns, the subscription handler is removed. Thus the client is
// able to process all messages sent by the server before the subscription
// handler is removed.
//
// Draining is particularly valuable with queue subscriptions preventing
// messages from being lost.
let c1 = 0;
const sid1 = nc.subscribe(subj, {queue: 'q1'}, () => {
c1++;
if(c1 === 1) {
nc1.drainSubscription(sid1, () => {
// subscription drained - possible arguments are an error or
// the sid (number) and subject identifying the drained
// subscription
});
}
});
// It is possible to drain a connection, draining a connection:
// - drains all subscriptions
// - after calling drain it is impossible to make subscriptions or requests
// - when all subscriptions are drained, it is impossible to publish
// messages and drained connection is closed.
// - finally, the callback handler is called (with possibly an error).
let c2 = 0;
nc.subscribe(subj, {queue: 'q1'}, () => {
c2++;
if(c2 === 1) {
nc1.drain(() => {
// connection drained - possible arguments is an error
// connection is closed by the time this function is
// called.
});
}
});
var nats = require('nats');
var fs = require('fs');
// Simple TLS connect
var nc = nats.connect({port: TLSPORT, tls: true});
// Overriding and not verifying the server
var tlsOptions = {
rejectUnauthorized: false,
};
var nc = nats.connect({port: TLSPORT, tls: tlsOptions});
// nc.stream.authorized will be false
// Use a specified CA for self-signed server certificates
var tlsOptions = {
ca: [ fs.readFileSync('./test/certs/ca.pem') ]
};
var nc = nats.connect({port: TLSPORT, tls: tlsOptions});
// nc.stream.authorized should be true
// Use a client certificate if the server requires
var tlsOptions = {
key: fs.readFileSync('./test/certs/client-key.pem'),
cert: fs.readFileSync('./test/certs/client-cert.pem'),
ca: [ fs.readFileSync('./test/certs/ca.pem') ]
};
var nc = nats.connect({port: TLSPORT, tls: tlsOptions});
See examples for more usage.
// Simple connect using credentials file. This loads JWT and signing key
// each time that NATS connects.
var nc = NATS.connect('connect.ngs.global', NATS.creds("./myid.creds"));
// Setting nkey and signing callback directly.
var nc = NATS.connect(url, {
nkey: 'UAH42UG6PV552P5SWLWTBP3H3S5BHAVCO2IEKEXUANJXR75J63RQ5WM6',
sigCB: function(nonce) {
return sk.sign(nonce);
}
});
// Setting user JWT statically.
var nc = NATS.connect(url, {
userJWT: myJWT,
sigCB: function(nonce) {
return sk.sign(nonce);
}
});
// Having user JWT be a function that returns the JWT. Can be useful for
// loading a new JWT.
var nc = NATS.connect(url, {
userJWT: function() {
return myJWT;
},
sigCB: function(nonce) {
return sk.sign(nonce);
}
});
// Connect with username and password in the url
var nc = NATS.connect("nats://foo:bar@localhost:4222");
// Connect with username and password inside object
var nc = NATS.connect({'url':"nats://localhost:4222", 'user':'foo', 'pass':'bar'});
// Connect with token in url
var nc = NATS.connect("nats://mytoken@localhost:4222");
// Connect with token inside object
var nc = NATS.connect({'url':"nats://localhost:4222", 'token':'mytoken'});
// Publish with closure, callback fires when server has processed the message
nats.publish('foo', 'You done?', function() {
console.log('msg processed!');
});
// Flush connection to server, callback fires when all messages have
// been processed.
nats.flush(function() {
console.log('All clear!');
});
// If you want to make sure NATS yields during the processing
// of messages, you can use an option to specify a yieldTime in ms.
// During the processing of the inbound stream, we will yield if we
// spend more then yieldTime milliseconds processing.
var nc = nats.connect({port: PORT, yieldTime: 10});
// Timeouts for subscriptions
var sid = nats.subscribe('foo', function() {
received += 1;
});
// Timeout unless a certain number of messages have been received
// the callback for the timeout. The callback for the timeout
// provides one argument, the subscription id (sid) for the
// subscription. This allows a generic callback to identify
// where the timeout triggered.
nats.timeout(sid, timeout_ms, expected, function() {
timeout = true;
});
// Auto-unsubscribe after MAX_WANTED messages received
var sid = nats.subscribe('foo', {'max':MAX_WANTED});
nats.unsubscribe(sid, MAX_WANTED);
// Multiple connections
var nats = require('nats');
var nc1 = nats.connect();
var nc2 = nats.connect();
nc1.subscribe('foo');
nc2.publish('foo');
// Encodings
// By default messages received will be decoded using UTF8. To change that,
// set the encoding option on the connection.
nc = nats.connect({'servers':servers, 'encoding': 'ascii'});
// PreserveBuffers
// To prevent payload conversion from a Buffer to a string, set the
// preserveBuffers option to true. Message payload return will be a Buffer.
nc = nats.connect({'preserveBuffers': true});
// Reconnect Attempts and Time between reconnects
// By default a NATS connection will try to reconnect to a server 10 times
// waiting 2 seconds between reconnect attempts. If the maximum number of
// retries is reached, the client will close the connection.
// To change the default behaviour specify the max number of connection
// attempts in `maxReconnectAttempts` (set to -1 to retry forever), and the
// time in milliseconds between reconnects in `reconnectTimeWait`.
nc = nats.connect({'maxReconnectAttempts': -1, 'reconnectTimeWait': 250});
// emitted whenever there's an error. if you don't implement at least
// the error handler, your program will crash if an error is emitted.
nc.on('error', function(err) {
console.log(err);
});
// connect callback provides a reference to the connection as an argument
nc.on('connect', function(nc) {
console.log('connected');
});
// emitted whenever the client disconnects from a server
nc.on('disconnect', function() {
console.log('disconnect');
});
// emitted whenever the client is attempting to reconnect
nc.on('reconnecting', function() {
console.log('reconnecting');
});
// emitted whenever the client reconnects
// reconnect callback provides a reference to the connection as an argument
nc.on('reconnect', function(nc) {
console.log('reconnect');
});
// emitted when the connection is closed - once a connection is closed
// the client has to create a new connection.
nc.on('close', function() {
console.log('close');
});
// emitted whenever the client unsubscribes
nc.on('unsubscribe', function(sid, subject) {
console.log("unsubscribed subscription", sid, "for subject", subject);
});
// emitted whenever the server returns a permission error for
// a publish/subscription for the current user. This sort of error
// means that the client cannot subscribe and/or publish/request
// on the specific subject
nc.on("permission_error", function(err) {
console.error("got a permissions error", err.message);
});
See examples and benchmarks for more information.
The following is the list of connection options and default values.
Option | Aliases | Default | Description |
---|---|---|---|
encoding |
"utf8" |
Encoding specified by the client to encode/decode data | |
json |
false |
If true, message payloads are converted to/from JSON | |
maxPingOut |
2 |
Max number of pings the client will allow unanswered before raising a stale connection error | |
maxReconnectAttempts |
10 |
Sets the maximum number of reconnect attempts. The value of -1 specifies no limit |
|
name |
client |
Optional client name | |
nkey |
`` | See NKeys/User Credentials | |
noEcho |
false |
Subscriptions receive messages published by the client. Requires server support (1.2.0). If set to true, and the server does not support the feature, an error with code NO_ECHO_NOT_SUPPORTED is emitted, and the connection is aborted. Note that it is possible for this error to be emitted on reconnect when the server reconnects to a server that does not support the feature. |
|
noRandomize |
dontRandomize , NoRandomize |
false |
If set, the order of user-specified servers is randomized. |
pass |
password |
Sets the password for a connection | |
pedantic |
false |
Turns on strict subject format checks | |
pingInterval |
120000 |
Number of milliseconds between client-sent pings | |
preserveBuffers |
false |
If true, data for a message is returned as Buffer | |
reconnect |
true |
If false server will not attempt reconnecting | |
reconnectTimeWait |
2000 |
If disconnected, the client will wait the specified number of milliseconds between reconnect attempts | |
servers |
urls |
Array of connection url s |
|
sigCB |
`` | See NKeys/User Credentials | |
tls |
secure |
false |
This property can be a boolean or an Object. If true the client requires a TLS connection. If false a non-tls connection is required. The value can also be an object specifying TLS certificate data. The properties ca , key , cert should contain the certificate file data. ca should be provided for self-signed certificates. key and cert are required for client provided certificates. rejectUnauthorized if true validates server's credentials |
token |
Sets a authorization token for a connection | ||
url |
uri |
"nats://localhost:4222" |
Connection url |
useOldRequestStyle |
false |
If set to true calls to request() and requestOne() will create an inbox subscription per call. |
|
user |
Sets the username for a connection | ||
usercreds |
`` | See NKeys/User Credentials. Set with NATS.creds() . |
|
userjwt |
`` | See NKeys/User Credentials | |
verbose |
false |
Turns on +OK protocol acknowledgements |
|
waitOnFirstConnect |
false |
If true the server will fall back to a reconnect mode if it fails its first connection attempt. |
|
yieldTime |
If set, processing will yield at least the specified number of milliseconds to IO callbacks before processing inbound messages |
Our support policy for Nodejs versions follows Nodejs release support. We will support and build node-nats on even-numbered Nodejs versions that are current or in LTS.
To run the tests, you need to have a nats-server
executable in your path. Refer to the server installation guide in the NATS.io documentation. With that in place, you can run npm test
to run all tests.
Unless otherwise noted, the NATS source files are distributed under the Apache Version 2.0 license found in the LICENSE file.