System Design FAQ: Top Questions

A Real-Time Chat System enables instant messaging between users, with delivery guarantees, presence indicators, channel-based grouping, and history storage.

📋 Functional Requirements

• 1:1 and group messaging
• Online/offline presence indicators
• Typing indicators and delivery status
• Persistent message history and search

📦 Non-Functional Requirements

• Low latency (<100ms message delivery)
• Fault tolerance (messages not lost)
• Scalability for millions of concurrent users

🏗️ Core Components

• WebSocket Gateway: Persistent connections for bidirectional communication
• Message Broker: Kafka or Redis pub/sub for routing
• Chat Service: Handles send, receive, and typing events
• Storage: PostgreSQL/MongoDB for chat history

🔌 WebSocket Server Example (Node.js)

const WebSocket = require('ws');
const wss = new WebSocket.Server({ port: 8080 });

wss.on('connection', function (ws) {
  ws.on('message', function (message) {
    console.log('received: %s', message);
    // Echo back to client or forward to Kafka
    ws.send(`Received: ${message}`);
  });
});
        

📦 Kafka Chat Topic Design

topic: chat.room.{room_id}

message format:
{
  "sender": "user123",
  "room_id": "eng-discussion",
  "text": "Hello team!",
  "timestamp": 1718223201
}
        

📖 Chat History Schema (MongoDB)

{
  "_id": ObjectId("..."),
  "room_id": "product-launch",
  "messages": [
    { "sender": "u1", "text": "Welcome", "sent_at": ISODate("2025-06-11T10:01:00Z") },
    { "sender": "u2", "text": "Thanks!", "sent_at": ISODate("2025-06-11T10:02:00Z") }
  ]
}
        

🟢 Presence Tracking

• Use Redis or etcd to store current active users
• Expiry key TTL to simulate timeouts
• Send heartbeats from frontend

🧪 Message Delivery Guarantee

• Client sends message → gateway → Kafka → chat processor → store + notify peers
• ACK sent to sender after storage success

📈 Observability

• Message send/receive delay
• WebSocket open/close connection rate
• Message drop count / backlog in broker

🧰 Tools/Infra Used

• WebSocket: ws (Node.js), socket.io, Golang net/http
• Broker: Kafka, Redis Streams
• Storage: PostgreSQL, MongoDB

📌 Final Insight

Real-time chat demands careful orchestration of connections, storage, delivery paths, and observability. Prioritize delivery guarantees, online presence accuracy, and sub-second latency for the best UX.