I am currently using CMS's in multiple projects. However,
the data does not change constantly and therefore the
json-response of the CMS can be cached.
Why should I not permanently use the CMS-Server? Option 1: Having the CMS-Server on standby causes annoying initial load times. Options 2: Constantly running the CMS-Server is very pricey.
I have implemented a http-trigger in the CMS. Every time the data changes, this cloud function gets called. The Function fetches a bunch of CMS-routes and stores them in a google storage bucket.
You can set the cached routes with environment variables
in the .env file.
Be sure to disable caching (either on the google storage bucket - can be quite annoying to set up - or on the fetching side).
With this I only have to run the CMS-Server when I am working
on the data and can even trigger the build-cms-data-script
manually and control the data rebuild by disabling the
auto-rebuild hook.
I also got the initial load times from >7s to <1s :)
You can find scripts for deploying, setting CORS policies
and making all files inside a bucket puplic in the directory
bash-scripts.