Provide additional context on authentication failure

When attempting to connect to a 6.6.0 or 7.0 cluster with bad credentials, and then performing cluster.WaitUntilReady, gocb attempts to retry until timeout, at which point it returns a TimeoutError. The TimeoutError doesn't provide any information to indicate the underlying problem was authentication - the inner error is just the timeout, and RetryReasons only contains 'NOT READY'.

Sync Gateway needs to communicate back to users that the root cause of the failed connection was authentication-related.

In addition, I'd be interested to know whether auth-related failure should be treated as fail-fast (no retry), or if the SDK is intentionally retrying on auth failure to handle the case provisioning scenarios where RBAC setup hasn't yet completed.