We should add an end to end tutorial for developer usage - taking an AWS RDS resource as a dependency and covering:

• Configuration
• Costs
• Security Checks
• Consumption via environment variables

Currently if i terminate a namespace, the resources that need to be deleted - if they fail on deletion, then it will get stuck in terminating mode and i am unable to clean up the cloud resource.

There are two issues:

1. The controller logs can't get displayed even if it can be successful on the cleanup because it can't schedule the pod to get logs as the namespace is in terminating mode
2. The controller shouldn't let me terminate a namespace if there are things i need to do in the cloud first or if i need to perform additional tasks like clean up resources in an s3 bucket or change the module so a force can be done etc. Whatever the reason, i don't want to be left with cloud resources that are now detached from my cluster and start to increase cloud costs. It would be nice to block some how the termination of the namespace if things fail in teranetes and pass that as a flag --block-ns-deletion

As terranates is using jobs under the hood, when the job has run, obviously it finishes but then there is no status after a period of time to retrieve i.e. the job has finished, even if failed and hence we are unable to follow the logs of the job to see the failure and why it failed at the time of execution.

It would be good if the pods actually remained to get logs from or if there was a nice way to be able to see the full log output of what happened somewhere, so you can get the status and the error of that status etc to be able to fix things up.

We should include a page covering the points/frequency at which the terranetes controller accesses the Cloud providers (AWS, GCP, Azure), if it's only on user action (CRUD on resources), or other periodic background jobs (i.e. drift detection), and how these can be controlled (directing to docs on how to disable drift detection). Similarly for cost estimates, which external service(s) it calls and at what times / intervals.

The documentation recommends creating a secret with a key of SSH_KEY_AUTH. I found that that key doesn't work. I had to dig into the code for the controller to find that it is looking for a key of SSH_AUTH_KEY

This line in the docs needs to be updated to use the right variable:

Add a tutorial in the admin section on how to configure the terraform-controller to use Vault.

The FAQs currently live under the tf2helm docs, but should be generalised to cover both products and a shared page across both docs spaces. Extend the docs with a feature comparison against other operators in this space.

We need a section on prometheus metrics which are exposed - and have it auto generate it ..

We should add a tutorial in the developer section of how to use FluxCD (GitRepository / HelmChart etc) as a source for the terraform module.

The source would be: http://source-controller.<namespace>.svc.cluster.local/<asset>