Skip to content
System Design Lab

Free · Interactive · Bilingual

System design you can poke, break and replay.

Stop memorising boxes and arrows. Every lesson is a live, deterministic simulation: inject a failure, turn a knob, step through each message and watch the metrics react.

Loading simulation…

Live: inject a failure and watch the circuit open.

Three ways in

Featured patterns

View all

Scaling lessons

View all

Real-world problems

View all

How interactive learning works

  1. 01

    See the architecture

    Each lesson starts from a diagram of real components — services, brokers, caches, databases — that you can zoom into and inspect.

  2. 02

    Run the scenario

    A deterministic discrete-event engine simulates every request. Same seed, same result, so you can replay and compare.

  3. 03

    Break it on purpose

    Take a dependency down, slow it, crash a consumer. Metrics come from simulated events, never from made-up numbers.

  4. 04

    Understand the trade-off

    Switch to advanced mode for percentiles, tuning parameters, production notes and Java/Spring code.

Built on one engine. Every simulation on this site reuses the same behaviors — client, service, circuit breaker, retry, timeout, cache, load balancer, Kafka topic, queue, saga — wired together declaratively.

Learning paths

Reliable event-driven systems

From the dual-write problem to exactly-once effects with outbox and idempotent consumers.

7 lessons

  1. 1Queue-Based Load Leveling
  2. 2Competing Consumers
  3. 3Dead Letter Queue
  4. 4Transactional Outbox
  5. 5Idempotent Consumer
  6. 6Customers Charged Twice: Duplicate Kafka Processing
  7. 7Saga

Keep data consistent across services

Own your data, then keep it correct without distributed locks: outbox, sagas, CQRS and event sourcing.

7 lessons

  1. 1Database per Service
  2. 2Saga
  3. 3Transactional Outbox
  4. 4Orders Without Invoices: Lost Events After Commit
  5. 5CQRS
  6. 6Event Sourcing
  7. 7One Crash Freezes Checkout: Two-Phase Commit Gone Wrong

Run everywhere, all the time

Stateless instances, autoscaling and capacity, then CDNs, failover and multiple regions.

6 lessons

  1. 1Stateless Application Design
  2. 2Autoscaling
  3. 3Capacity Planning
  4. 4Content Delivery Network (CDN)
  5. 5High Availability
  6. 6Multi-Region Architecture

Shape and observe a system

Monolith or microservices, where to draw boundaries, and the metrics, logs and traces that show what is happening.

6 lessons

  1. 1Modular Monolith
  2. 2Microservices Architecture
  3. 3Service Decomposition
  4. 4Metrics and Monitoring
  5. 5Centralized Logging
  6. 6Distributed Tracing