Production Debug Under Access Constraints: Fixing a Live 500 Without Dashboard Access

A live admin endpoint was returning HTTP 500 in production. The constraint made it interesting: we had no access to the hosting dashboard and no access to their error tracking — and on a first engagement with an unfamiliar vendor, we should not have had either.

We installed error tracking into the client's monorepo, then diagnosed the fault from the repository, the captured error data, and client-pasted log lines alone. We asked only for environment variable names — never their values. The fix landed inside a two-task fixed scope, and the instrumentation stayed behind so the next failure reports itself instead of needing a debugger.