Serverless

FaaS execution model: functions triggered by events, billed per-invocation, no server management — cold starts are the main trade-off.

Why we need this / what value this brings

No idle cost, no server patching/scaling to manage — ideal when load is bursty or infrequent.

When to use this

Short-lived, event-triggered, stateless work — not for long-running connections or workloads sensitive to cold-start latency.

How to use or implement this

Keep functions small and stateless; move anything needing persistent connections (DB pools) to a model that reuses them across invocations if cold starts become a problem.

Research questions

• Cloudflare Workers' cold-start story vs AWS Lambda's — relevant to both the FullStackFusions BYOK chat and SAA-C03 study.

---

Empty folder — drop notes, links, and findings here as you research.