We built boxkite because lots of ML drift monitoring solutions are too complex and heavyweight: this is a pure python library exposing a pure prometheus endpoint, then we implemented KL divergence in one massive PromQL query :-)
Looks interesting and installation steps seem to be straightforward, will definitely try it out in my stack!
Thanks, would love to know how you get on, please join #boxkite-monitoring on MLOps.community slack to let us know how it works out for you :-)
Nice one! super happy to see this and I'm looking forward to going through the demo.