Skip to content
DecisionNerd Skills

Compute — Cloudflare

Cloudflare: Pages, Workers, durable objects/KV/R2 as needed — edge-first compute and static hosting.

Fit

Section titled “Fit”
  • Global latency, edge logic, static+API at the rim
  • Workers-sized workloads (CPU/memory/time limits OK)
  • Overlap with Serverless; choose Cloudflare when edge network is the product constraint

Skip when

Section titled “Skip when”
  • Long jobs, heavyweight native deps, or regional data residency that forces a classic cloud region — Clouds / Docker

Starter DAG

Section titled “Starter DAG”
flowchart TD
  A[Worker or Pages project] --> B[Local wrangler/dev smoke]
  B --> C[repos ci: test + deploy preview]
  C --> D[ESC for upstream API secrets]
  D --> E[Promote production route]
  E --> F[observe-it: worker errors, latency, quota]

Skills

Section titled “Skills”

repos ci · repos secrets · test-it · observe-it · kiss · document-it (limits & binding names)

Watchouts

Section titled “Watchouts”
  • Ignoring isolate CPU limits until prod
  • Mixing Pages and three Workers with no ownership map
  • Duplicating the same site on Vercel and Cloudflare without a reason