revisit log levels

Logs are quite noisy.

We currently log all internal K8s requests to `INFO`. We should probably downgrade these to `DEBUG` to reduce noise. I think it would be nice to see:
- `ERROR` - the Devnet API can't handle something and is broken - should prompt a response
- `WARN` - a user's Devnet instance encounters a problem, or there's an invalid request made
- `INFO` - a user makes a request to the devnet (just one log/request)
- `DEBUG` - the internal requests that the Devnet API makes to the k8s API

No files or tests are named. Start by locating the Devnet API's logging for internal Kubernetes requests and user requests, then trace how errors and invalid requests are classified. Done means the four stated levels are applied consistently, with internal Kubernetes traffic at DEBUG and one INFO entry per user request.

Scritto dal modello di indicizzazione a partire dal testo della issue.