Design & Fix

A live system is failing. Find the cause, change the design and run it again until it holds.

Build & Test your own Design any system and stress-test it with spikes and failures. Free once a week, unlimited on paid plans
45 incidents
Status Incident Difficulty Inspired by Stars Actions
Live incident L3A celebrity post overloads one cache Medium Instagram —
Live incident L3Fleet migration: half the rides land on the old small host Medium Uber —
Pro Start
Live incident L3Match start: the fantasy leaderboard melts the database Medium Dream11 —
Pro Start
Live incident L3Writing suggestions depend on one AI provider, and it's at its limit Medium Grammarly —
Pro Start
Live incident L3Scrapers find the internal service URLs Medium X (Twitter) —
Pro Start
Live incident L3Portfolio refreshes slow down the trades they're trying to show Medium Groww —
Pro Start
Live incident L3A TV feature sends search and bestseller traffic to one database Medium Etsy —
Pro Start
Live incident L3Workspace sign-up waits on two slow partners Medium Databricks —
Live incident L3Shorts uploads time out while transcoding Medium YouTube —
Pro Start
Live incident L3Dinner rush: order writes overwhelm the database Medium Zomato —
Live incident L4Ad click tracking drowns the database Hard Snap —
Live incident L4One Postgres is the queue, the file store and the search engine Hard Notion —
Live incident L4Lunch rush: orders time out and get lost Hard DoorDash —
Live incident L4A redesign put a cache and a queue in all the wrong places Hard LinkedIn —
Live incident L4Premiere night: an app cache can't stream video Hard Netflix —
Pro Start
Live incident L5Big Billion Days checkout: three things are wrong at once Hard Flipkart —
Pro Start
Live incident L5Last-over finish: the live score page goes down Hard Disney+ Hotstar —
Pro Start
Live incident L5Checkout waits on "complete the look" Hard Myntra —
Pro Start
Live incident L5Opening a busy channel reads history through a queue Hard Discord —
Pro Start
Live incident L5Fleet telemetry: chained load balancers and a dedupe cache Hard Tesla —
Pro Start