We have a data lake with our commits at work. I use it to look at developer activity data. Not for performance reasons but to see which of our code bases are active or not. I look at things like iteration time or length a PR/CR sits, number of revisions, etc. This lets me see where we have opportunities to speed up development loops for engineers. Also lets me understand if a service that is having poor performance or growing in cost is under active development. I'll find outliers and go dig into them. This also lets me justify work on tooling because I can quantify the problems and thus the benefits.