-
Notifications
You must be signed in to change notification settings - Fork 110
Description
Hi,
I was able to successfully launch it but had some difficulties along the way.
Maybe something worth highlighting in your docs :
- The version of terraform & terragrunt should be aligned. For terraform 0.15, needs to be at least 0.29
Otherwise, Terraform can't pass properly commands to terraform
I had 0.14.x and matching terragrunt. I've switched the terraform using tfenv to 0.15. But terragrunt I've forgot , as there was nothing on its version management.
- As mentioned in docs, I've deleted secret manager.tf file. But, this blocked me further.
Got tons of errors like this.
So, I had to try solution number 2 : Creating secret with Json data, but with Dummy passwords
Error: Reference to undeclared local value │ │ on eks-kube-prometheus-stack.tf line 16, in locals: │ 16: gitlab_client_id = local.grafana_gitlab_client_id │ │ A local value with the name "grafana_gitlab_client_id" has not been │ declared.
- When I've deployed the solution, I wanted to log in to Grafana
There were no instructions how to get a password
I've seen that it is in the terragrunt output , but as it is a sensitive value it was "redacted". So, it is masked.
So, solution for this :
kubectl get secret --namespace monitoring kube-prometheus-stack-grafana -o jsonpath="{.data.admin-password}" | base64 --decode ; echo
- After it was deployed, now was time to destroy.
There were no destroy commands for terragrunt
Did the following :
cd layer2-k8s
terragrunt destroy
cd ../layer1-aws
terragrunt destroy
The order I guess is super important, as if not starting from k8s which provisions LB, this will block layer-1-aws from cleaning up VPC due to dependency on LB, SG, Subnet, Network interfaces. And, we will have to clean up it all manually
Other than that, went ok
Question:
- How to generate those client secrets and etc for auth through Gitlab?
Instructions for this will be nice to have (for all params u are using there)