-
Notifications
You must be signed in to change notification settings - Fork 3.4k
policy: No-op Identity Allocator #35973
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
Conversation
341bd73
to
db8efe6
Compare
/test |
Clustermesh failures seems like bug introduced in this PR:
corresponds to:
|
I see. Interesting. I only changed from struct pointer (*RemoteCache) to interface (RemoteIDCache). |
Yes, you can call method on nil pointer, but not on nil interface. On nil interface it causes panic. |
ce39b1d
to
edd01a8
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
thanks, lgtm for clustermesh
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
(requesting changes instead of comment, so this doesn't automatically get merged without comments being addressed)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks! One more comment re. (*Allocator).WatchRemoteKVStore
, otherwise LGTM.
**Description** As part of the network policy modularization and decoupling of the network policy enforcement system from the endpoint, the no-op allocator aims to provide a possibility of disabling the identity allocation system when network policies are disabled. **Motivation** The scalability of network policy feature is limited when the scalability dimensions, # of nodes and pod churn rate, are stretched. Cilium has many networking features that scale better, or are unaffected by these scalability dimensions. Currently, disabling network policies to reach higher scale for other features is the only feasible solution. Note: The plan is to improve scalability of network policies, so that it can reach higher scales that are required. **Follow-up** Subsequently, the plan is to agree on a configuration set that would determine if the identity allocation system should be disabled, and put the no-op allocator to use. Proposed config: - EnablePolicy=never - DisableCiliumEndpointCRD=true - EnableK8sNetworkPolicy=false - EnableCiliumNetworkPolicy=false - EnableCiliumClusterwideNetworkPolicy=false Signed-off-by: Dorde Lapcevic <dordel@google.com>
/test |
Ref: #33360
Description
As part of the network policy modularization and decoupling of the network policy enforcement system from the endpoint, the no-op allocator aims to provide a possibility of disabling the identity allocation system when network policies are disabled.
Motivation
The scalability of network policy feature is limited when the scalability dimensions, # of nodes and pod churn rate, are stretched. Cilium has many networking features that scale better, or are unaffected by these scalability dimensions. Currently, disabling network policies to reach higher scale for other features is the only feasible solution.
Note: The plan is to improve scalability of network policies, so that it can reach higher scales that are required.
Follow-up
Subsequently, the plan is to agree on a configuration set that would determine if the identity allocation system should be disabled, and put the no-op allocator to use.
Proposed config:
Signed-off-by: Dorde Lapcevic <dordel@google.com>