[beats receivers] Include exporter metrics and a way to determine when the collector is experiencing output backpressure in diagnostics

[cmacknz]

Today in diagnostics the Beat 30 second metric reports include output and queue metrics that allow seeing trends in performance and errors. In particular the queue metrics are frequently used to quickly determine if the output source is unable to keep up with data from the Beats.

Include a way to get equivalent information when the beat is translated into a receiver using an exporter instead.

Ideas to consider are at least:

• Including a snapshot of current metrics using the elasticdiagnostics extension
• Streaming internal telemetry to a file.
• Subscribing to internal telemetry and logging it or emitting interesting events when they are detected.

[elasticmachine]

Pinging @elastic/elastic-agent-control-plane (Team:Elastic-Agent-Control-Plane)

[cmacknz]

The simplest path is likely to just log a snapshot of the internal telemetry metrics every 30s or so like we do now in the new monitoring receiver that is coming.

We can iterate on this later if we want something nicer.