Issues

1. The memory usage and open handles charts should split on the component.id instead of the elastic_agent.process. The elastic_agent.process is not unique. For example there can be multiple filebeat processes. We want the split to be same as it is for the CPU Usage chart.
2. The queue depth/percentage charts are broken. They don't work at all. I think it is because they are looking at logs-* instead of metrics-*.
3. The cgroup memory chart could be more clear. Remove the "split by" (or only show elastic-agent) because all processes are a member of the same cgroup. This will make it easier to read the chart to see how close the value is to the limit. The "limit" line should also be made thicker because it's hard to notice (like 2px instead of 1px).
4. Nothing on this page indices when CPU throttling is occurring. Add a chart showing the system.process.cgroup.cpu.stats.throttled.ns metric. See https://github.com/user-attachments/assets/894e6e76-d4a1-4cc2-8171-6599170510d7.

This is an example visualization show the CPU throttling data: