2. Design a Distributed Rate Limiter
Enforce tenant quotas across servers while making burst behavior and outage policy explicit.
ClassicMedium
Guided practice
Start with a template. Work through each step. Ask Coach when you need a second opinion.
All designs 28
Ordered by design numberCompany tags are community-reported. Counts on cards show how many people reported that design.
Filtered designsClear filters
3. Design a Distributed Cache
Partition cached data, survive node loss and keep cache misses from overwhelming the source of truth.
ClassicMedium10. Design an API Gateway
Route API traffic with authentication, tenant limits and safe configuration rollout.
ClassicMedium11. Design a Container Orchestrator
Reconcile desired workloads, schedule containers and recover from failed nodes safely.
ClassicHard- Practice briefStart with your own approach
12. Roll out configuration to a disconnected fleet
Safely converge 40,000 gateways while some sites are disconnected and old controllers can return.
Infrastructure & SREHard - Practice briefStart with your own approach
13. Keep a regional ingress service alive
Survive a regional outage without confusing a health check with safe recovery.
Infrastructure & SREHard 29. Cache-Aside (Lazy Loading)
A read-heavy service sends every request to a database that is slower and more expensive per read than memory, and the same few rows are asked for over and over.
Infrastructure & SREMedium32. Token Bucket Rate Limiter
A limit enforced per server is not a limit: ten servers each allowing a hundred requests a minute allow a thousand. And a counter per fixed window lets twice the limit through across a window boundary.
Infrastructure & SREMedium34. Delay Queue Scheduler
Millions of future actions — reminders, retries, expiries — must fire near their due time. Scanning everything every minute does not scale, and an in-process timer dies with the process.
Infrastructure & SREMedium42. Design a Metrics Monitoring Platform
Collect 5 million samples a second from 500,000 hosts, store them as time series, chart them and page people.
ClassicMedium45. Design a Job Scheduler
Run 10,000 jobs a second within two seconds of their time, at least once, with retries, fairness between tenants and exactly-once effects.
ClassicMedium63. Design Real-Time Updates for Millions of Clients
Every goal on 25 million open apps within two seconds: SSE or WebSockets, pub/sub to the right server, resume by seq, hot topics, reconnect storms and push.
ClassicMedium66. Design for Read Scale: A Product Page Read a Million Times a Second
Serve a product page a million times a second: a cheap query, replicas, Redis and the edge, with hot keys, stampedes and invalidation handled.
ClassicMedium67. Design for Write Scale: A Million Writes a Second
A live TV vote at a million writes a second: spread by key, buffer in Kafka, aggregate the hot counter in two stages, shed what can wait.
ClassicMedium68. Design Large File Uploads and Downloads
Files up to 50 GB up and down without a byte through the API: presigned URLs, resumable multipart, S3 events, scanning and CloudFront.
ClassicMedium69. Design Long-Running Tasks: Async Jobs with Progress and Results
Accept in milliseconds, work in the background: leases and heartbeats, retries and dead letters, progress by SSE and webhooks, fairness across tenants.
ClassicMedium70. Design an In-Memory Data Store like Redis
An in-memory data store like Redis, from one event loop to Redis Cluster to a durable platform.
ClassicMedium73. Design a Durable Workflow Engine like Temporal
Workflows as code that survive any crash: event histories and replay, activities retried under timeouts, durable timers, sharded history.
ClassicMedium78. Design a Coordination Service like ZooKeeper
A coordination service like ZooKeeper: znodes, sessions and watches, a ZAB quorum of five with observers, and the recipes for elections and locks.
ClassicMedium79. Design Object Storage like Amazon S3
An S3-inspired object store: bytes before metadata, a partitioned namespace with cache coherence, then repair, heat, lifecycle and safe change. Three interview boards with explicit assumptions and AWS source boundaries.
ClassicMedium80. Design a Message Queue like Amazon SQS
Build a queue around visibility leases and explicit acknowledgement, then add replicated partitions, FIFO groups, tenant fairness and bounded redrive: three self-contained interview boards.
ClassicMedium86. Design an IP and URL Denylist
Deny requests by caller address, range or URL at a global edge, with lists from governments, threat feeds and detection reaching 300 proxies in seconds.
ClassicMedium91. Design a CI/CD System like GitHub Actions
Turn 10 pushes a second into job graphs on ephemeral runners, with fair queues per pool, leases fenced by attempt and live logs.
ClassicMedium97. Design a Distributed Logging and Error Tracking System
Ten terabytes of logs an hour from host agents through Kafka into tiered OpenSearch and S3, with regex search, live tail and exceptions grouped into issues.
ClassicMedium100. Design a Distributed File System
Billions of files, hundreds of petabytes, one strongly consistent tree: a namespace partitioned by directory, chunk servers, replication, repair, erasure coding. One-hour boards for junior, senior and staff, with the theory behind them.
ClassicMedium102. Design a GPU Cluster Scheduler for Training Jobs
Thousands of GPUs shared by many teams: gang admission, topology-aware placement, quotas with lending, checkpointed preemption and failure recovery.
ClassicMedium103. Design Model Weight Distribution to a GPU Fleet
A 500 GB model on 1,000 GPU servers in minutes: chunks and hashes, a topology-aware swarm, signed manifests, bandwidth budgets and waves.
ClassicMedium117. Design a Self-Healing GPU Fleet
Detect, classify and repair failing GPU nodes (Xid errors, ECC, NVLink and EFA faults, stragglers, silent corruption) without draining the fleet.
ClassicMedium