Add more logging to CPU profiling instrumentation and test on ECE

[rudolf]

During a recent SDH we tried to capture a CPU profile by passing the --profiler.signal argument and sending the signal but nothing happened (it's entirely possible this was a human error due to trying something new).

It would be useful if our profiling instrumentation had more logging to give some visibility:

• A log entry to say that profiling is enabled and which signal we're listening for
• A log entry when profiling is started
• A log entry when profiling is stopped including the filepath we're writing to before writing. Once written log the filepath again.

See https://github.com/elastic/kibana/blob/main/src/cli/profiler/profiler.js#L15

We can test this on serverless with:
https://github.com/elastic/kibana-team/blob/98cb8fce811479b283981b8cef170af528fa427a/docs/serverless/troubleshooting.mdx#cpu-profiling

We can try to follow steps similar to https://github.com/elastic/sdh-kibana/issues/5161#issuecomment-2569905689 for testing on ECE (would require help from an SRE)

[elasticmachine]

Pinging @elastic/kibana-core (Team:Core)