bench: support ledger team on a component-level benchmark

[deepfire]

Problem

Component-level ledger benchmarks are not reflective of the real-world performance on AWS.

Essence

We want to help ledger team to obtain a benchmark that would reflect the regression we see in AWS.

[deepfire]

This might be a relevant effort: https://github.com/input-output-hk/ouroboros-network/tree/nfrisby/CAD-4214-forge-times-via-db-analyser

cc @JaredCorduan @lehins