Procedure building images for testing and release purposes should be thersame

Even we want to build official image just before push on demand in both cases build should be done the same way. Right now haf_block_explorer and balance_tracker uses different scripts and jobs to do it, what due to bug in: https://gitlab.syncad.com/hive/common-ci-configuration/-/blob/develop/templates/docker_image_jobs.gitlab-ci.yml?ref_type=heads#L44 which hidden important part of before_script inherited from: https://gitlab.syncad.com/hive/common-ci-configuration/-/blob/develop/templates/docker_image_jobs.gitlab-ci.yml?ref_type=heads#L9 lead to image build failure.

All or most of code required to build the image should be stored in local scripts, to allow easy local reproducing of potential problems and make code much more readable.

If our image builder needs to be preconfigured somehow, we should extend base image for that (which should be used in all cases) instead of "patching" code by missing statements put into before_script handlers what is hard to maintain and understand.