Why should development and production use separate storage buckets?

Why should development and production use separate storage buckets?

Separate buckets help prevent development activity from affecting production photos. Developers need to test uploads, invalid files, and deletion, while users expect their wedding memories to remain available. Isolating the environments gives those activities distinct storage boundaries.

Prevent test work from touching real photos

A test cleanup script may remove every object associated with a fixture. If development and production share a bucket and broad credentials, a configuration mistake can reach real data.

Use clearly named environment-specific buckets and test images. Keep the related databases separate too, so development metadata cannot reference production objects accidentally. Bucket separation is helpful only when configuration and permissions preserve the distinction.

Apply separate credentials and policies

Give the development application credentials scoped to its own bucket. Production credentials should not be needed for routine local work. R2's authentication documentation describes bucket-scoped access.

Set each bucket's browser-origin policy for the environments that actually use it. Production should not inherit a permissive development configuration by accident. Also review cleanup rules, retention requirements, and public-access settings independently.

Preserve realistic testing without copying private data

Match important behavior across environments while using controlled fixtures. Test the same upload workflow, image validation, and deletion recovery that production uses. A separate bucket should not become an excuse for an unrealistic development setup.

Before releasing the gallery, confirm the deployed environment references the correct bucket and database together. Exercise a controlled production upload if the release process allows it, then clean up that specific test item. Record which service configuration belongs to each environment. This makes maintenance safer and helps an AI agent continue work without guessing where test operations are allowed to run.