Reconcillation issue in s3 controller

[mandeepgoyat]

Describe the bug
I noticed that reconciliation feature not working continuously in s3 controller(v0.1.0)
I am expecting that once s3 controller installed then it should detect the drift continuously and restore the resource back to its original state.
Reconciliation happens but only when I delete the controller's pod. Looks like not working continuously like Crossplane

Steps to reproduce

1. Installed the s3 controller via helm
2. Create IAM role for service account
3. Create s3 resource via manifest file
4. s3 bucket created successfully
5. If i delete the s3 bucket manually(via AWS console)
6. s3 bucket not created automatically

Expected outcome

1. S3 controller should detect the drift continuously and re create the s3 bucket automatically .

Environment

• Kubernetes version --1.23
• Using EKS (yes/no), ---Yes, version is 1.22
• AWS service targeted --S3

[RedbackThomson]

Hey Mandeep.
We have common logic in all of the ACK controllers that when the resource reaches its desired state, it will pause reconciliation for a while. It would be CPU intensive to continuously check for drift at a relatively short interval, so we leave that to the K8s API server default requeue time, which I think is 8 hours. I understand that waiting 8 hours for drift detection is probably too long to rely on. From what I can see, Crossplane also defaults back to the API server for their resources (see code here). Perhaps you could tell me what experience you've had with drift detection, and what an appropriate amount of time would be?