TODO.md

Custom images with dependencies installed on:

Solution is to build the images locally and push it to a registry. After that you can set the base_image in the pipeline codes and use it. HACK: You also can connect to minikubes docker daemon and build the image there. After that k8s would have that image locally and no need to pull it. Caveat: Pulling images from a remote repository may be a threat because the container should be able to communicate with the outside world. Check it out

Run this http://192.168.100.11:8080/arka/arka_ai/-/blob/main/ml_classification.py?ref_type=heads in pipeline

Resource and namespace management