Logical layers #22
Description
A random idea that would be amazing to see down the line: different views/layers to look at construct relationships in different contexts.
For example, you could have a “Networking” view, where you would only see the constructs that are placed in Security Groups, you’d see which constructs are in which VPCs, their relationships (SG rules), etc.
Then you could do an “IAM layer” that would for example show which constructs have IAM access to other resources (explicitly granted with CDK’s grant* to make this easier, maybe?).
Or a “Computing” view, where you’d see all of your long-running compute resources and maybe their relationships (ASG groups / ECS clusters / Services / containers)
This reflects my biggest issue with similar tools - they often don’t represent the way we reason about our resources, especially when we start looking at bigger apps. And I don’t think it’s possible to do that with a single view (maybe it is, but I can’t come up with one).
Maybe you could leverage tags at first (implement a convention that devs could utilize to explicitly express logical construct relationships / groups). Dunno.