Production-only Failure

A production-only failure is a bug that appears after deploy even though local development and local production build seem fine.

The common mistake is treating it as one class of problem. It usually comes from a small set of mismatches:

Environment variables: missing, renamed, injected at build time instead of runtime, or exposed only to server code.
Runtime differences: Node version, edge runtime, serverless timeout, filesystem access, region, memory, or cold start.
Origin differences: CORS, cookie domain, callback URL, redirect URI, or mixed protocol.
Data differences: production database state, permissions, migrations, seed data, or stale cache.
Build differences: dead-code elimination, bundled public variables, image paths, basePath, or generated routes.

A useful bug note should include exact URL, deploy id, runtime, request id, environment name, expected response, actual response, and one confirmed difference from local.

The goal is to stop saying “works locally” and start naming the boundary where the environment changed.

Production only failure

// COMMENTS

ON THIS PAGE

Production only failure

// COMMENTS ↓ Newest First

ON THIS PAGE

// COMMENTS