Opening your workspace…
Opening your workspace…
A connected path from capacity basics to reasoning under failure.
Predict, run the model, inspect the trace, and defend a trade-off.
Find the bottleneck during a traffic spike, test a warm cache, then challenge the same fix with cold traffic.
Use a token bucket to bound admitted traffic, then distinguish deliberate rejection from serving every request.
Builds on capacity cacheCombine workload estimates, cache assumptions and admission control, then defend the remaining design decisions.
Builds on capacity cache, rate limiterAdd a node to a seeded hash ring, inspect key movement, then reveal why a hot key still overloads one owner.
Builds on capacity cacheRead behind an asynchronous replica, change the read route, then expose the availability cost of a regional failure.
Builds on capacity cacheCompare timestamps alone with coordinated and worker/sequence strategies, then challenge duplicate worker assignment and clock assumptions.
Builds on url shortenerReplay a post-effect, pre-acknowledgement boundary, observe duplicate effects, and separate idempotency from queue capacity.
Builds on capacity cacheHandle transient notification failures with bounded retries and idempotency, then see why a permanent failure still needs an explicit terminal state.
Builds on queues workersCompare push, pull and threshold-based hybrid fanout while tracing write work, dropped updates and read amplification.
Builds on queues workers, capacity cacheCompare lost messages, replay and duplicate suppression across a disconnect using explicit message identities and a bounded buffer.
Builds on queues workersTrace an uncertain timeout and a later provider effect, suppress duplicate charges and reconcile an outcome the caller cannot yet prove.
Builds on queues workersUse event and queue evidence to separate normal capacity from a dependency slowdown, then inspect retry amplification and recovery.
Builds on replication consistency, queues workersAll scenarios use original teaching material and explicit synthetic assumptions. Technical review covers the declared model; it does not establish real infrastructure performance or interview readiness.