After we upgrade a process we want to be able to understand if the process is healthy or not, I don't think for the first release we need something really complicated like trying to interrogate the outputs.

I think it could be a simple as the process is started, logs works, the web API is started and it's ready to execute configuration. Maybe it's even a flag that is set when the beats have completely started and are about to run the configuration.

I would like to start having a conversation with the Services team to define what it could be.

Also, we would like to be able to retrieve that state through in our manager implementation.

cc @urso @andresrc