[`hydrate-orchestrator`] Improve error logging

[juanjosevazquezgil]

Motivation

While this module currently shows all errors it finds, it does it in an unintuitive way, at least when it's a Helm error. The error is displayed, but Dagger continues outputing its own logs and then adds a helm exited with process code 1 message to the end of the action logs, so any operator trying to debug the actual helm error needs to scroll up and look for the true helm logs, which can be a bit of a pain given the volume of logs outputed by Dagger.

We want to fix this problem however possible: maybe we could add an ::error:: tag to the relevant logs, maybe we could upload them to an artifact and attach it to the workflow action. Whichever option we decide to go with, we have to me mindful of multiple errors or warnings happening in the same execution, and handle them accordingly, as Dagger doesn't seem to stop its execution until it's forced to do so (i.e., we've had cases where an error happened, the module continued execution and reported another different error when it terminated)

Acceptance criteria

• Establish a fix for the problem
• Implement the fix
• Test the fix when multiple warnings/errors happen in the same execution