Performance analysis of integration tests

[dylan-smith]

It would be good if we could collect some data from how long each of our integration tests take to run over time.

It would be good to know if the performance is improving / degrading over time of a full end-to-end migration, and how much variability there is.

I'm hoping we can extract this data from the build logs (the verbose logs are all timestamped) and put it all somewhere we can query / report on it (e.g. kusto).