All tracks
Track · 0/13 done
Consumer apps at scale
What breaks when millions of people use one app at once: pushes, launches, sales, retries and bad rollouts.
- 01Lesson · 2 minPush notification stampede
- 02Read · Real-time & chat · every client reconnects at onceReconnect storms
- 03Lesson · 2 minVirtual waiting room
- 04Lesson · 2 minSeat holds
- 05Read · Databases · two buyers, one last itemOptimistic transactions
- 06Read · Payments · the double chargeIdempotency keys
- 07Lesson · 2 minAutoscaling lag
- 08Lesson · 2 minServerless cold starts
- 09Lesson · 2 minGraceful degradation
- 10Lesson · 2 minFan-out on write vs read
- 11Lesson · 2 minRequest coalescing
- 12Read · Databases academy · one busy key on one nodePartitioning and sharding
- 13Read · Consensus · "I just saved it, where is it?"Replication lag
- 14Lesson · 2 minThe self-repair loop
- 15Read · Deployments · the global config pushStaged config rollouts
- 16Lesson · 2 minDNS resolver overload
- 17Lesson · 2 minThe dependency everything shares
- 18Lesson · 2 minOut-of-band access
- 19Read · Consensus · a network split, two primariesSplit brain & fencing tokens
- 20Lesson · 2 minCell-based architecture
- 21Read · DNS & HTTP · shrinking one customer's blast radiusShuffle sharding