Consider capturing logs of processes orchestrated by Elastic Agent

Today, we capture Elastic Agents logs just like we do for any other Pod. Often, for issues related to the Agent, it is needed to look at log output of processes (Beats) managed by the Agent to discover issues. These logs are stored in the Pod. If we could include them in the diag, it would be much easier to investigate issues related to, for example, connecting to output and certificate validation.