Log a warning that file-configs override service-discovery configs

[sodabrew]

What does this PR do?

Logs a warning if a service-discovery configured check will be ignored if the same-named check was already configured from a file.

Motivation

I have an haproxy.yaml file pointing at one instance of haproxy that I want to monitor on all servers. I am using Docker service discovery to configure other containers for service checks. This week I wanted to deploy a new use case for an haproxy container, and so I diligently added the Docker labels for an haproxy check. Nothing happened. Radio silence in the logs files.

Testing Guidelines

An overview on testing
is available in our contribution guidelines.

Additional Notes

A better fix would be to merge the instances list from the file with the service discovered-instances list. Ironically, there is code in the service discovery config module to do exactly this, and then it is dutifully discarded if a file config exists. But first I want to solve the head-banging-on-wall problem of failing to log the discarded config.

[sodabrew]

Do Agent 6 or 7 allow checks defined in multiple sources per my use case?