03/05/2021, 7:21 AM
How did you deal with the chicken-and-egg problem of two resources depending on each other and creation ordering problem? Or did you not need it?


03/05/2021, 11:39 PM
The way I was planning to solve this one ATM is to use stack references. I'm building my own config schema to define my "stack", but am separating regional resources from edge ones. Then I plan to use stack reference to existing resources, based on their names, not IDs. I'll then look them up as I'm creating/defining the Front Door (or AppGW) rules, etc.
I built my Cosmos database schemas this way, so all I need to do is update a config file with a new index and push a PR
my reasoning behind this approach is that I want to create my resources and verify them independently before rotating them into production using front door. This would mean two commits, one to create the stack resources (and verify they're working), another to enable them in Front Door (or AppGW)
it also means that Front door isn't the "owner" of them (ie. creating them), merely references them.