Scalability & Performance Patterns

The default vs stateless architecture decision.

How to split work across nodes without creating new bottlenecks.

UX tradeoffs, timeout design, and how to pick the right model.

Token bucket, leaky bucket, and where to enforce limits.

What happens when consumers cannot keep up — and how to handle it.

Threads, processes, and event loops — and when parallelism actually helps.

Fast for one user vs fast for a million — they require different solutions.

Isolating workloads so one overloaded path does not bring down the rest.

Deciding which work gets resources first when capacity is constrained.

Reducing per-request overhead by grouping work intelligently.

When caching is a performance tool vs when it masks a deeper problem.

Scaling up before you need to — and not paying for it when you do not.

Absorbing burst load without pre-provisioning for peak at all times.