-
Notifications
You must be signed in to change notification settings - Fork 879
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add the Ability to Specify Namespace of Scheduler Estimator Services for Karmada Scheduler #5448
Comments
Yeah, I think it's a reasonable use case. What do you think? @whitewindmills @Garrybest |
+1 |
/assign |
Nice catch. I think it's useful. |
@whitewindmills We talked about this issue at the community meeting yesterday. One question from Joe (@jabellard) is do you want to take this task as he sees the assignment you made on this? |
@whitewindmills , I already looked into how to implement this change and have a PR I want to push. |
okay, you can continue it. |
@RainbowMango , just pushed a PR for this. |
What would you like to be added:
When the scheduler estimator feature is enabled for the Karmada scheduler component, we can then provision scheduler estimator instances for member clusters that will provide the scheduler with more accurate information regarding resources for those member clusters.
To gain that information, given some member cluster, the scheduler needs to contact the service for that member cluster's scheduler estimator deployment. As of today, it uses a naming convention for discovering those services. By default, given some member cluster
m1
, it will look for a service namedkarmada-scheduler-estimator-m1
in thekarmada-system
namespace.The service name prefix is configurable, but the namespace is not. This feature request is for providing the ability to configure the namespace in which the scheduler should try to discover the service.
Why is this needed:
At Bloomberg, we're currently building a managed Karmada platform and want to use the Karmada operator to manage the entire lifecycle of managed Karmada instances. Each tenant's control plane will live in a namespace in our host/management cluster. As such, we would also like to keep scheduler estimator instances scoped to their respective tenancy's namespace to avoid polluting the
karmada-system
namespace with instances across all managed tenancies.The text was updated successfully, but these errors were encountered: