We coukd also keep that manually, unless we want to have Atomic builds in this file. Adding it to the makefile could help, as we do for translation source files.

We could then do something like:

make releases

This could create a file as actually in /static and we can just commit and push it to the repo. Would be a sort of half-automation process. Again, the same as we have for POT files with destination zanata.

I added the instruction make releases to our makefile to run the script and create the json file directly in the right directory. Web members can just run it, commit and push.
We cannot do more ATM, I'll add it also to the docs for updating the websites, in order to run it at every release (Alpha, Beta and GA). I will mention also to update the list of releases within the script itself.
See rev 4a28738

@pfrields is there anything else we can do now as websites team to make this process smoother?

We maintain many other scripts before every release and run them once we merge them into master.
After 2 full release cycles I feel the process is rather straightforward and has become just part of our release update process.

• Update the getfedora.org/build/releases-json.py script
• Go to /getfedora.org directory
• Run make releases and commit the new created releases.json file

I'm going to close this as fixed, but if you feel there are better ways or have ideas about how to make it more effective, please reopen or file a new issue.