We don't capture the second RV draw yet. Unfortunately, I don't know if it's the best use of my time to capture that as well.
We really don't want our startup tests to subtlely change so I thought it was worth writing some quick tests.
As mentioned in the class comment, we could use an alternative implementation to measure perf from outside the application however it's more complicated to set up and less consistent with our cold startup methodology so we'll start with this.