Discovery / PoC for provisioning multiple copies staging using terraform workspaces
I believe we can use workspaces inside environment/staging to create multiple copies of staging, allowing us to provision and tear down staging environments easily for testing.
Some items off the top of my head:
-
Come up with a DNS conventions if we are really going to have disposable staging endpoints. -
Decide if we want to do any repo copying from snapshots, an example of this can be seen in the GEO setup. -
It probably doesn't make sense to do the db restore in the terraform run if it takes 6h, what is the best way to automate it? -
Will this impact takeoff? -
Decide whether we should move to shared modules for what we have in staging as-is, my vote is yes but this may not play well with workspaces. Maybe we can figure out a way to make it work? -
Decide how to handle monitoring and logging. We probably won't be standing up more than one a week, keep one environment on a dedicated dashboard? -
Investigate using db snapshots instead of s3 db restore for faster provisioning -
Automate / script the sanitizing process -
Demo the provisioning
The output of this should be a decision on whether this is a good idea, and if so a list of tasks for making it happen.
Edited by John Jarvis