Support generic health check about tf-controller HOT 3 OPEN

This will require the controller image to have cli/tools installed to run. TerraformRunner can help with this by allowing users to customize their runner image (feature?).

from tf-controller.

We'll decide about this one later, as it might compromise our security.

from tf-controller.

Thinking a bit more about this, I think an event-driven approach might be a good way to approach generic health-checks. For example, in the context of AWS I can subscribe to an event-bridge bus and listen to resource creation events; the flux notification controller could be configured to receive these events and forward to the terraform controller. This approach allows us to avoid any cloud provider specific implementations from the terraform controller.

We still need some logic around mapping events to terraform outputs, defining the health check and tracking it's state. The flux notification controller doesn't have a receiver for event-bridge or pub/sub but this is something we could add.

from tf-controller.