Not sure, if I'm going in the wrong direction here - so let me know if there is another way to achieve this :-)
We have two different cluster types (hyperscaler) and our tests need to be executed with also two different permission sets (admin and customer).
Based on the documentation, I solved the hyperscaler "issue" with labels:
chainsaw test --selector "test-target in (all, h1)"...
---
apiVersion: chainsaw.kyverno.io/v1alpha1
kind: Test
metadata:
name: my-test
labels:
test-target: all # all, h1, h2
(...)
The two different permissions would be handled with the cluster configuration. \
$KUBECONFIG_PATH will be the same file with one cluster but two different users and contexts.
chainsaw test \
--cluster customer=$KUBECONFIG_PATH:customer \
--cluster admin=$KUBECONFIG_PATH:admin \
--no-cluster
There is now the use case, that certain tests should be executed in the customer and administrator contexts. \
At the moment, I see the only option to duplicate the test and adjusting the cluster: customer to cluster: admin because cluster is a string and not a list.
Describe your question
Not sure, if I'm going in the wrong direction here - so let me know if there is another way to achieve this :-)
We have two different cluster types (hyperscaler) and our tests need to be executed with also two different permission sets (admin and customer).
Based on the documentation, I solved the hyperscaler "issue" with labels:
The two different permissions would be handled with the
cluster
configuration. \$KUBECONFIG_PATH
will be the same file with one cluster but two different users and contexts.There is now the use case, that certain tests should be executed in the
customer
andadministrator
contexts. \ At the moment, I see the only option to duplicate the test and adjusting thecluster: customer
tocluster: admin
becausecluster
is a string and not a list.Are there any suggestions on how to handle this?
chainsaw version Version
v0.2.11
Additional context
No response