Rate limiting (token bucket / leaky bucket)

We package an app with its dependencies into an image, which runs as a container - identical on your laptop and in production. "Works on my machine" stops being an excuse.

Continuous Integration / Delivery: every commit is automatically built, tested and (optionally) deployed. This pipeline lets us ship safely many times a day, without manual mistakes.

We run two identical environments: blue is live, green is the new version. Once green is verified we flip traffic to it; on trouble we flip back instantly. Zero-downtime releases with instant rollback.

We add more machines/instances (scale out) instead of one bigger box (vertical, scale up). For stateless services this wins: cheaper, more elastic, no ceiling. State goes to a separate store.

Distributes incoming traffic across multiple instances - the front door that gives you redundancy and smooth scaling. Health checks remove dead instances, so one failure stays invisible to users.

We follow one request across every service using a trace ID (e.g. OpenTelemetry). In a microservices system this is how we pinpoint which service slowed down or failed - no guessing.