Is python mojo prediction supposed to be so slow?

Question

I have an H2O python application that I recently converted to containers for kubernetes. As part of the conversion I started saving MOJO files when training and using MOJO p