Make service config more explicit

Currently when a service config is supplied, we do not know if things are configured correctly as we desired. Service configs are applied silently. This makes it hard to know what goes wrong when using a service config. For example, in #5724, a retry service config is supplied by user but does not correctly applied to the service since the service name in the config was wrong. User sees nothing but retry is not happening. Is it possible to give out some information about what's going on when there is a service config?

Start by tracing how service configs are applied in the Java gRPC implementation, using the retry configuration problem described in #5724 as the concrete case. Determine where a supplied service name fails to match and what diagnostic information is currently available. Done means users receive useful information when a service config is invalid or not applied.

Vom Indexierungsmodell aus dem Issue-Text verfasst.