04/19/2022, 11:43 AM
Question is if the dependency from the helm chart’s perspective would be better if the image was referred to using the target address rather than the final image name as seen by docker? Much in the same way is done for supporting depdendencies between docker image targets too, where the
instruction may refer to the upstream image using its target address paired with a build arg, like this:
Copy code
ARG upstream=src/upstream:image-address
FROM $upstream