Discuss/Fix missing eks coverage in functional tests

[nithyatsu]

Investigate potential missing eks coverage in functional tests.
#8033 adds dapr to radius installation, but this change requires an eks cluster to be created by following dapr docs. The tests did not catch the scenario. This is not a radius bug.
But we have to decide whether we want our tests to cover basic functionality on all of kind, aks and eks and add tests based on the decision.

AB#13818

[brooke-hamilton]

@nithyatsu please provide more context on this. We may need an additional team discussion.

[radius-triage-bot]

👋 @nithyatsu we need more information.

To help us investigate further, we need additional information to reproduce or understand this issue. Please ensure you have steps to reproduce and information about your OS, rad CLI version, and runtime version.

A project maintainer will reply on this issue soon to ask for additional specifics if needed.

For more information on our triage process please visit our triage overview

[kachawla]

@nithyatsu please provide more context on this. We may need an additional team discussion.

@nithyatsu @vishwahiremat can you add more detailed information here about impact and the fix?

[vishwahiremat]

@nithyatsu please provide more context on this. We may need an additional team discussion.

@nithyatsu @vishwahiremat can you add more detailed information here about impact and the fix?

@kachawla here are the steps followed manually to get the rad install kubernetes command working with eks cluster, which needs to be added as part of the workflow for eshop-aws.

1. Cluster Creation creation step: EKS cluster should be created using the configurations mentioned here
2. Get the security group: Use this command to get the security groups
   aws eks describe-cluster --name <cluster-name> --query 'cluster.resourcesVpcConfig.clusterSecurityGroupId' --region <region>
3. Add Dapr requirements for sidecar access and default storage class:

aws ec2 authorize-security-group-ingress --region [your_aws_region] \
--group-id [your_security_group] \
--protocol tcp \
--port 4000 \
--source-group [your_security_group]

kubectl patch storageclass gp2 -p '{"metadata": {"annotations":{"storageclass.kubernetes.io/is-default-class":"true"}}}'

[nithyatsu]

Background: release 0.41 ran into issues in Samples validation. rad init was failing on eks cluster. We traced it back to dapr support we added in radius 0.41. The issue was not caught since we do not run functional tests in eks, aks as part of PR workflows. As part of PIR, we decided to log a story that captures potential ways to avoid such issues. #8199 was created following the PIR.
This issue was logged when I first found the issue while working on release. It is a duplicate and can be closed.

[kachawla]

Background: release 0.41 ran into issues in Samples validation. rad init was failing on eks cluster. We traced it back to dapr support we added in radius 0.41. The issue was not caught since we do not run functional tests in eks, aks as part of PR workflows. As part of PIR, we decided to log a story that captures potential ways to avoid such issues. #8199 was created following the PIR. This issue was logged when I first found the issue while working on release. It is a duplicate and can be closed.

@nithyatsu Can you describe the immediate impact this bug has? My understanding is that it is causing sample workflows to fail consistently along with the manual steps needed for every release. If this is correct then we need a fix to address the immediate pain given that #8199 is more longer term.

@kachawla , yes those are the 2 impacts. For immediate relief we have to update the samples workflow like Vishwa has mentioned.

For users who wants to use radius on eks, eks must be created following the dapr guidelines. Other than that there are no changes.

The impact on team is that

• samples will consistently fail on eks
• in every release, on call engineer should create eks manually using the guidelines and test eshop manually on it.

[kachawla]

@nithyatsu please provide more context on this. We may need an additional team discussion.

@nithyatsu @vishwahiremat can you add more detailed information here about impact and the fix?

@kachawla here are the steps followed manually to get the rad install kubernetes command working with eks cluster, which needs to be added as part of the workflow for eshop-aws.

1. Cluster Creation creation step: EKS cluster should be created using the configurations mentioned here
2. Get the security group: Use this command to get the security groups
   aws eks describe-cluster --name <cluster-name> --query 'cluster.resourcesVpcConfig.clusterSecurityGroupId' --region <region>
3. Add Dapr requirements for sidecar access and default storage class:

aws ec2 authorize-security-group-ingress --region [your_aws_region] \
--group-id [your_security_group] \
--protocol tcp \
--port 4000 \
--source-group [your_security_group]

kubectl patch storageclass gp2 -p '{"metadata": {"annotations":{"storageclass.kubernetes.io/is-default-class":"true"}}}'

Thank you @vishwahiremat!

[kachawla]

As discussed in the handoff discussion yesterday, marking this as important since it is causing consistent sample failures and has been causing operational pain during release. Immediate scope is to fix the test samples workflow. @nicolejms @vishwahiremat

[radius-triage-bot]

We've prioritized work on this issue. Please subscribe to this issue for notifications, we'll provide updates as we make progress.

We also welcome community contributions! If you would like to pick this item up sooner and submit a pull request, please visit our contribution guidelines and assign this to yourself by commenting "/assign" on this issue.

For more information on our triage process please visit our triage overview