Change the numbers, then watch what happens to every request. Start with the whole architecture in 3D, then go deep on networking, caching, load balancing, queues, rate limiting and sharding.
A typical web backend, from the user to the database. Drag to orbit, pinch or scroll to zoom, and tap any component to learn what it does.
DNS lookup, TCP handshake, TLS, then the request itself. Each arrow is a packet crossing the network, so distance multiplies everything.
Cache-aside pattern: the API checks Redis first and only queries the database on a miss, then stores the answer for a while (the TTL).
Each server has its own speed, like real machines on shared hardware. Overload one and it starts rejecting requests. Crash one and see how long it takes health checks to notice.
Producers drop messages into a queue and move on. Consumers pull them off one at a time. Failed messages are retried, and after too many failures they go to a dead-letter queue.
A rate limiter protects an API by rejecting requests over a limit with HTTP 429 Too Many Requests. Different algorithms treat bursts very differently.
Seven short steps, each one changing the picture so you can see the idea instead of just reading it. Use Next to move on, or tap the dots to jump.