[metrics] gather host-level metrics during testing

[bitwiseguy]

We should track host-level metrics such as memory usage, disk usage, cpu usage, etc and combine that with the existing block processing time data replayor is already collecting to give a more holistic view of how the execution client is performing. The following metrics are provided by geth and would be of interest:

• Memory usage: Tracks memory allocations, frees, and total usage
• CPU usage: Measures CPU utilization by Geth and the overall system
• Disk usage: Monitors disk read/write operations and chaindata size
• Database operations: Measures read/write operations to the chaindata
• Database size: Tracks the size of the chaindata directory
• Block processing: Measures block import times and rates

Consider integrating with go flight recorder but would need to consider how to support reth / other clients that don't use golang.

Should make this data format compatible with Grafana

[GrapeBaBa]

@bitwiseguy There are some uncertainties. It appears that node_exporter can be used to gather host metrics. Prometheus combined with Grafana is suitable for visualizing client-specific metrics. The Go flight recorder serves as a trace/profile tool for pinpointing program performance bottlenecks, which doesn't directly relate to metrics. So, what is the reasoning behind integrating it?