Open alexppg opened 3 years ago
This would make using CI much easier, rather than having a long command line full of flags.
@alexppg, @day1118, and @zegl
I have been playing around with this, since I agree a configuration file would be useful. My current local implementation is to generate the configuration file on-demand (kube-score mkconfig). The configuration file directives could be overridden at command line, preserving existing functionality/behavior. For example, one could still exclude a test using the --ignore-test flag without editing .kube-score.yml.
At present, the configuration looks as such --
$ ./kube-score mkconfig $ cat ./kube-score.yml (### BTW I know this isn't actually valid YAML yet. I'm still trying to settle on what it should look like ###)
object: CronJob checks:
object: Deployment checks:
object: HorizontalPodAutoscaler checks:
object: Ingress checks:
object: NetworkPolicy checks:
object: Pod checks:
object: PodDisruptionBudget checks:
object: Service checks:
object: StatefulSet checks:
object: all checks:
Is this in line with what you were envisioning as a user?
Hi @kmarteaux - This is great news!
I had assumed the syntax would be as simple as matching the cli flags. This would make it very easy to follow and reason about:
disableIgnoreChecksAnnotations: false
enableOptionalTest:
- test1
- test2
ignoreTest:
- test3
- test4
I do think the syntax that you have proposed feels like overkill, but either way, I think it would be more natural to replace status: "on"
with enabled: true
Thanks
Anthony, thanks for the feedback. That is why I asked, it felt like I was overthinking it. I will adjust accordingly.
I agree with @day1118. It seems a pretty understandable approach. Thanks @kmarteaux !
It would be awesome to have the possibility to have a .kube-score.yml file in which we could configure whatever we need. I think at the time it would mostly be tests to ignore, but it may be interesting for the future.
This might serve as inspiration: https://docs.kubelinter.io/#/configuring-kubelinter
PD: Thanks for your work, kube-score is awesome!