Injected Atlas Configuration

[qcaas-nhs-sjt]

The broadsea implementation currently creates it's own custom version of the atlas configuration file, this is then injected into the container and environment variables are replaced. This is similar to how it is done in Atlas' own docker image configuration, however it's injecting different files to do this than in the core atlas repository.

I would argue that this is bad practice as:

• We're promoting injecting code directly into the container in an uncontrolled manner which isn't ideal from a security point of view
• Any new configuration added to Atlas will not be reflected in the broadsea implementation unless we then make the same changes to broadsea
• This duplicates effort without any real gain.

Have created a PR in the atlas repository OHDSI/Atlas#2917 which makes the majority of configuration available via config currently using the same methodology as used here. If this gets implemented then we should seek to remove the injected files from broadsea and use the ones in the resulting container image.

[leeevans]

@qcaas-nhs-sjt you make some very valid points.

@alondhe what do you think? I would be in favor of making this change. This would be a breaking change for folks that have previously installed Broadsea, so it would have to be well communicated and documented.

[leeevans]

@qcaas-nhs-sjt I'd appreciate it if you could create a separate PR to upload your helm chart in this repo.

It would be great to have an initial version that we can iterate on over time. If you agree, you could create a sub-directory tree something like below and add it there.

github repo root dir
      - k8s
         - atlas