Improve container registry error message with bad AWS permissions
Can we improve the error messaging when pushing to container registry with bad AWS permissions?
Ran into an AWS permissions issue with a customer on an S3 backed registry. IAM / Bucket Policy permission were incorrect.
Docker login worked fine, but docker push would loop through retires until it errored out with a 503.
curl -v localhost:5001/debug/health also showed a 503 with no explanation of what the error was.
Customer ticket -> https://gitlab.zendesk.com/agent/tickets/89057 (internal only)