When the stack is being re-created to generate a new version of the qgreenland package, we have some downtime while we wait for nginx to come back up
AC: Split NGINX and Luigi in to two stacks with their own responsibilities: Build and Serve. Only bring down the "build" stack when deploying new code, NGINX should have as little downtime as possible.
Could we achieve this by pushing qgreenland docker images to dockerhub?
When the stack is being re-created to generate a new version of the qgreenland package, we have some downtime while we wait for nginx to come back up
AC: Split NGINX and Luigi in to two stacks with their own responsibilities: Build and Serve. Only bring down the "build" stack when deploying new code, NGINX should have as little downtime as possible. Could we achieve this by pushing qgreenland docker images to dockerhub?