Événements envoyés par le serveur pour une diffusion unidirectionnelle
Diffusez des mises à jour en direct aux clients avec le décorateur @Sse et les observables RxJS.
Événements envoyés par le serveur pour une diffusion unidirectionnelle est une leçon NestJS Enterprise Backend APIs gratuite sur CoddyKit. Ceci est la leçon 3 sur 4. Tu peux lire la leçon complète ci-dessous gratuitement — puis la pratiquer en direct dans le navigateur avec un éditeur de code intégré et un tuteur IA 24/7. Elle fait partie du parcours d'apprentissage NestJS Enterprise Backend APIs, et ta progression se synchronise sur le web et l'application CoddyKit. Le cours NestJS Enterprise Backend APIs comprend 4 leçons au total.
Certaines parties de cette leçon n'ont pas encore été traduites et s'affichent en anglais.
Why Server-Sent Events?
Server-Sent Events (SSE) let a server push a continuous stream of updates to a client over a single, long-lived HTTP connection. It is the simplest way to deliver one-way, server-to-client live data such as notifications, progress bars, or dashboard metrics.
- One-way only: the server talks, the client listens. There is no client-to-server channel on the same connection.
- Plain HTTP: no special protocol upgrade like WebSockets need. It rides on a normal GET request.
- Auto-reconnect: the browser's
EventSourcereconnects automatically if the connection drops.
In NestJS, SSE is a first-class feature exposed through the @Sse() decorator combined with RxJS observables.
SSE vs WebSockets
Both stream live data, but they solve different problems. Choosing the right one is an architectural decision.
- SSE: server → client only, text-based, runs over HTTP/1.1 or HTTP/2, built-in reconnection and event IDs. Ideal for feeds, alerts, and progress.
- WebSockets: full-duplex (both directions), binary or text, requires a protocol upgrade. Ideal for chat, multiplayer games, and collaborative editing.
If your clients only need to receive updates, SSE is lighter, easier to scale behind standard HTTP infrastructure, and requires no extra client library. Reach for WebSockets only when the client must also push messages in real time.
The wire format
SSE is just a streaming HTTP response with the content type text/event-stream. Each message is a block of text fields separated by newlines, and each block ends with a blank line.
data:the payload (often a JSON string).event:a named event type the client can listen for.id:a message identifier used for reconnection viaLast-Event-ID.retry:the reconnection delay in milliseconds.
You rarely format this by hand in NestJS — the framework serializes a typed object into these fields for you — but knowing the shape helps you debug with curl.
// Raw text/event-stream bytes the server emits
// (NestJS builds this for you from a MessageEvent)
const sample =
'id: 42\n' +
'event: heartbeat\n' +
'data: {"status":"ok","ts":1718000000}\n' +
'\n';
process.stdout.write(sample);Your first @Sse endpoint
The @Sse() decorator marks a controller method as an SSE stream. Instead of returning a plain value, the method returns an RxJS Observable. Every value the observable emits becomes one SSE message sent to the client.
NestJS expects each emitted value to be a MessageEvent-shaped object with a data property. It automatically sets the text/event-stream headers and keeps the connection open.
import { Controller, Sse, MessageEvent } from '@nestjs/common';
import { interval, map, Observable } from 'rxjs';
@Controller('events')
export class EventsController {
@Sse('clock')
clock(): Observable<MessageEvent> {
return interval(1000).pipe(
map((n) => ({ data: { tick: n, time: new Date().toISOString() } })),
);
}
}The MessageEvent shape
NestJS exports a MessageEvent interface that maps directly onto the SSE wire fields. Only data is required; the rest are optional.
data— string or object. Objects are JSON-stringified automatically.type— becomes theevent:field (a named event).id— becomes theid:field, enabling resume-on-reconnect.retry— becomes theretry:field in milliseconds.
Returning a well-typed object keeps your stream self-documenting and lets clients subscribe to specific named events.
import { MessageEvent } from '@nestjs/common';
function buildEvent(orderId: string, n: number): MessageEvent {
return {
id: String(n),
type: 'order.updated',
retry: 5000,
data: { orderId, sequence: n },
};
}
console.log(buildEvent('ord_123', 7));Pushing domain events with a Subject
A fixed interval is fine for clocks, but real systems push when something happens. The idiomatic pattern is an RxJS Subject living in a service. Your business logic calls .next() on the subject whenever an event occurs, and the SSE endpoint simply exposes the subject as an observable.
This cleanly decouples the producer (any service) from the transport (the SSE controller).
import { Injectable, MessageEvent } from '@nestjs/common';
import { Subject, Observable } from 'rxjs';
@Injectable()
export class NotificationsService {
private readonly stream$ = new Subject<MessageEvent>();
emit(payload: unknown): void {
this.stream$.next({ type: 'notification', data: payload });
}
asObservable(): Observable<MessageEvent> {
return this.stream$.asObservable();
}
}Wiring the service to the controller
The controller injects the service and returns its observable from an @Sse() method. Any other part of the app — a queue consumer, a webhook handler, a cron job — can inject the same service and call emit() to broadcast to every connected client.
Because a plain Subject is multicast, all subscribers receive each emission. This is exactly what you want for a shared notification feed.
import { Controller, Post, Body, Sse, MessageEvent } from '@nestjs/common';
import { Observable } from 'rxjs';
import { NotificationsService } from './notifications.service';
@Controller('notifications')
export class NotificationsController {
constructor(private readonly notifications: NotificationsService) {}
@Sse('stream')
stream(): Observable<MessageEvent> {
return this.notifications.asObservable();
}
@Post()
publish(@Body() body: { message: string }) {
this.notifications.emit(body);
return { accepted: true };
}
}Per-user filtered streams
A global subject broadcasts to everyone. In an enterprise API you usually want each client to receive only their events. Use RxJS operators like filter and map to tailor the stream per request, reading the user from a route param or the authenticated request.
The @Sse() method can accept normal route decorators such as @Param() and @Req(), so you can scope the observable to the current user.
import { Controller, Param, Sse, MessageEvent } from '@nestjs/common';
import { Observable, filter, map } from 'rxjs';
import { EventsBus } from './events.bus';
@Controller('users')
export class UserFeedController {
constructor(private readonly bus: EventsBus) {}
@Sse(':userId/feed')
feed(@Param('userId') userId: string): Observable<MessageEvent> {
return this.bus.events$.pipe(
filter((e) => e.userId === userId),
map((e) => ({ type: e.kind, data: e.payload })),
);
}
}Heartbeats keep the connection alive
Proxies, load balancers, and browsers may close an idle connection. A heartbeat — a periodic comment or no-op event — keeps the pipe warm. With RxJS you merge your real event stream with a slow timer.
Send heartbeats as a distinct event type (or as SSE comment lines) so clients can ignore them. A common interval is every 15–30 seconds, comfortably under typical proxy idle timeouts.
import { merge, interval, map, Observable } from 'rxjs';
import { MessageEvent } from '@nestjs/common';
export function withHeartbeat(
source$: Observable<MessageEvent>,
): Observable<MessageEvent> {
const heartbeat$ = interval(15000).pipe(
map((): MessageEvent => ({ type: 'heartbeat', data: 'ping' })),
);
return merge(source$, heartbeat$);
}Cleanup, errors, and backpressure
When a client disconnects, NestJS unsubscribes from your observable. Make sure your stream releases resources on unsubscribe — use finalize() for cleanup and never leak timers or listeners.
- Use
catchErrorto convert errors into a final event instead of crashing the stream. - Use
finalizeto log or decrement a connection counter when the client leaves. - Beware backpressure: a fast producer with a slow client buffers in memory. Throttle or sample high-frequency sources.
import { Observable, catchError, finalize, of } from 'rxjs';
import { MessageEvent } from '@nestjs/common';
export function safeStream(
source$: Observable<MessageEvent>,
onClose: () => void,
): Observable<MessageEvent> {
return source$.pipe(
catchError((err) =>
of<MessageEvent>({ type: 'error', data: { message: err.message } }),
),
finalize(onClose),
);
}Consuming the stream from a client
Browsers consume SSE with the native EventSource API. It connects, dispatches messages, and reconnects automatically. Listen to the default message event for unnamed data, or add listeners for your named event types.
Note that EventSource only supports GET and cannot set custom headers, so auth is usually done via cookies or a token in the query string. Tools like curl -N are great for quick debugging from the terminal.
// Browser-side consumer
const es = new EventSource('/notifications/stream');
es.addEventListener('notification', (e) => {
const payload = JSON.parse(e.data);
console.log('new notification', payload);
});
es.addEventListener('heartbeat', () => {
// keep-alive, ignore
});
es.onerror = () => console.warn('reconnecting...');Quick Check
Test your understanding of when and how to use SSE in NestJS.
Recap
You learned how to stream one-way live updates from NestJS using Server-Sent Events:
- SSE is server-to-client only over plain HTTP, with built-in browser auto-reconnect — choose it over WebSockets when clients only need to receive.
- The
@Sse()decorator turns a controller method into a stream that returns anObservable<MessageEvent>; each emission becomes one message. - A
MessageEventmaps to the wire fieldsdata,type,id, andretry; objects are JSON-serialized for you. - Push real domain events with a shared
Subjectin a service, and scope per-user streams withfilterandmap. - Keep connections healthy with heartbeats via
merge, and clean up safely usingcatchErrorandfinalize. - Consume on the client with the native
EventSourceAPI.
Questions Fréquemment Posées
La leçon « Événements envoyés par le serveur pour une diffusion unidirectionnelle » est-elle gratuite ?
Oui — le texte complet de « Événements envoyés par le serveur pour une diffusion unidirectionnelle » est gratuit à lire ici sur le web. Pour la pratiquer de manière interactive (un éditeur de code intégré et un tuteur IA 24/7) et déverrouiller le reste du cours NestJS Enterprise Backend APIs, passe à CoddyKit PRO. Le cours NestJS Enterprise Backend APIs comprend 4 leçons au total.
Qu'est-ce que j'apprendrai dans « Événements envoyés par le serveur pour une diffusion unidirectionnelle » ?
Diffusez des mises à jour en direct aux clients avec le décorateur @Sse et les observables RxJS. Tu pratiques NestJS Enterprise Backend APIs avec du code pratique que tu exécutes directement dans le navigateur, et un tuteur IA 24/7 répond à tes questions au fur et à mesure que tu avances dans la leçon.
Dois-je avoir de l'expérience pour commencer NestJS Enterprise Backend APIs ?
Aucune expérience préalable n'est requise. NestJS Enterprise Backend APIs sur CoddyKit est structuré pour les débutants jusqu'aux apprenants avancés, donc tu peux commencer ici ou depuis le début et avancer à ton rythme. Ceci est la leçon 3 sur 4.
Combien de temps prend la leçon « Événements envoyés par le serveur pour une diffusion unidirectionnelle » ?
La plupart des leçons CoddyKit prennent environ 5–10 minutes. Chacune est courte et interactive, tu progresses régulièrement et tu repiques exactement où tu t'es arrêté sur le web et l'app.
Peux-tu écrire et exécuter du code dans cette leçon NestJS Enterprise Backend APIs ?
Oui. Chaque leçon NestJS Enterprise Backend APIs inclut un éditeur de code intégré, tu écris et exécutes du vrai code directement dans ton navigateur et tu reçois des retours IA instantanés — aucune configuration locale requise.
Toutes les leçons de ce cours
- Passerelles WebSocket avec l’adaptateur Socket.IO
- Authentifier et protéger les connexions Socket
- Événements envoyés par le serveur pour une diffusion unidirectionnelle
- Mettre à l’échelle le temps réel avec un adaptateur Redis Pub/Sub