Real-Time Apps with MongoDB Change Streams

Welcome to Day 3! ✨

In the old days, to know if data changed, we had to Poll the database every 5 seconds. setInterval(() => checkForUpdates(), 5000) This kills your server and database.

MongoDB introduced Change Streams, which allows your app to Subcribe to database changes in real-time.

1. How it works

It uses the MongoDB Oplog (Operations Log). When a write happens, MongoDB pushes an event to all subscribers.

2. Basic Example (Node.js)

const collection = db.collection('orders');
const changeStream = collection.watch();

changeStream.on('change', (next) => {
  console.log("Something changed!", next);
});

If you insert a document:

{ operationType: 'insert', fullDocument: { ... } }

If you update a document:

{ operationType: 'update', updateDescription: { updatedFields: { status: 'shipped' } } }

3. Filtering Changes

You don’t want to listen to every change. Maybe you only care when an order is “cancelled”.

const pipeline = [
  { $match: { 'fullDocument.status': 'cancelled' } }
];

const changeStream = collection.watch(pipeline);

4. Use Cases

1. Notifications: User A comments on a post -> Change Stream detects it -> Push Notification to User B.
2. Audit Logs: Stream all changes to a separate Audit collection or ElasticSearch.
3. Cache Invalidation: When data changes, clear the Redis cache key.

🧠 Daily Challenge

1. Open a terminal script that watches the users collection.
2. Open another terminal (or Compass) and manually update a user.
3. Watch your script print the update in real-time! ⚡

See you on Day 4 for TTL Indexes (Self-destructing data)! 💣

Next Step

Next: Ttl Indexes Lifecycle →