[Internal] Improve Logging to be more relevant - collect more info on what we don't expect and less of what we do

[don-de-marco]

Steam may send some notification events to us although we might not have explicitly asked for information. For example, when you stop playing a game, we'll receive this notification from the Steam backend, which we currently just ignore:

2023-06-27 00:27:00,833 - steam_network.protocol.protobuf_client - INFO - Processing message ServiceMethodResponse PlayerClient.NotifyLastPlayedTimes#1
2023-06-27 00:27:00,833 - steam_network.protocol.protobuf_client - WARNING - Unparsed message, no idea what it is. Tell me
2023-06-27 00:27:00,833 - steam_network.protocol.protobuf_client - WARNING - job name: "PlayerClient.NotifyLastPlayedTimes#1"

We're already handling some of them but I believe there were some changes and we should pay a bit of attention to them at some point. It's nothing that is broken or non-functional but something that we could use to notify Galaxy of changes ahead of their own polling mechanism.

[ABaumher]

On a related note: can/should we set the json-rpc log level to warning or higher so we can have smaller log files? I feel like a lot of the info calls don't tell us anything and make finding errors immensely more tedious. It would increase our reliance on the galaxy client log to track down some misbehaving code but we already need than anyway.