Compatible Images with Cloudera Machine Learning Platform

We are interested to deploy all of the Supabase components (kong, studio, etc) into our CML platform as applications. (Refer to [CML Runtime]( There are changes that we identify below (non exhaustive), in order for the images to be runable on top of cloudera 'runtime scheduler' (ref [Custom PBJ compatible runtime]( Potential changes: • Main process port inside container has to be 8100 (one of ports opened by CML as nodeport), optional: would be great if this can be set from the environment variable. Example: kong listens to 8000, however cloudera scheduler likely do similar to docker run --name <kong-container-name> --network <supabase-network> -p 8100:8100 <kong image> and that -p 8100:8100 can't be changed. • The entire image resource can be run/accessible as non-root. Reason: both cloudera and us are adding and symlinking addons to /usr/lib and other paths traditionally only accessible by root (we can identify just limited numbe…