Cleanup of error logs and categorization

[jeniawhite]

Motivation
We need to make sure that we prioritize and categorize errors that we are logging. This has an effect on the serverless deployment process. During the promotion of versions we rely on several metrics, slos, alerts. We query the error percentage as well in order to understand if we should fail the promotion. With the current state of cloudbeat sometimes there are errors that we aren't sure if should be logged and blocking the promotion os serverless versions.

Definition of done
What needs to be completed at the end of this task

• cloudbeat should have a logging strategy to avoid pollution of logs
• Should clean up the logs and align with serverless workflows to make sure that we aren't affecting the promotion process without actual problems

Out of scope
What should not be included in this task

Related tasks/epics

• https://github.com/elastic/agentless-controller/issues/268