How to use this guide

Read the cheat sheet first, then jump into the topics you are weak on. For interviews or design reviews, treat every section as a set of questions: what problem is this solving, what does it cost, what breaks, and how will we observe it?

• Start with requirements: traffic, latency, durability, consistency, security, tenant isolation, and budget.
• Prefer simple designs until there is evidence that complexity is justified.
• Every optimization has a failure mode. Caching introduces staleness; queues introduce duplicates; sharding introduces cross-shard complexity.
• For .NET examples, assume ASP.NET Core, EF Core, dependency injection, structured logging, and cancellation tokens.

Core principle

Scale the bottleneck you actually have. Do not add Redis, Kafka, sharding, or multi-region replication just because they appear in architecture diagrams.