Kubernetes v1.35 Introduces Watch-Based Route Reconciliation in Cloud Controller Manager
Kubernetes version 1.35 introduces a new alpha-stage feature gate named CloudControllerManagerWatchBasedRoutesReconciliation within the k8s.io/cloud-provider library. This update addresses inefficiencies in previous versions (up to v1.34), where the route controller in the Cloud Controller Manager (CCM) reconciled routes at fixed intervals, causing unnecessary API requests to cloud providers even when no changes occurred. The new mechanism leverages watch-based informers, similar to other controllers in the library, to trigger reconciliation only when specific events occur, such as node addition, deletion, or updates to podCIDRs and status addresses. Additionally, a random reconcile interval between 12 and 24 hours is performed to ensure consistency. This change aims to reduce API load without altering the core logic of the reconciliation loop, ensuring minimal impact on existing configurations. Developed by SIG Cloud Provider, this feature can be enabled via a specific feature gate flag. It represents a significant optimization for cloud-native infrastructure management, enhancing efficiency and reducing operational overhead for users managing Kubernetes clusters on various cloud platforms.
Wire timeline
Kubernetes v1.35 Introduces Watch-Based Route Reconciliation in Cloud Controller Manager
Kubernetes version 1.35 introduces a new alpha-stage feature gate named CloudControllerManagerWatchBasedRoutesReconciliation within the k8s.io/cloud-provider library. This update addresses inefficiencies in previous versions (up to v1.34), where the route controller in the Cloud Controller Manager (CCM) reconciled routes at fixed intervals, causing unnecessary API requests to cloud providers even when no changes occurred. The new mechanism leverages watch-based informers, similar to other controllers in the library, to trigger reconciliation only when specific events occur, such as node addition, deletion, or updates to podCIDRs and status addresses. Additionally, a random reconcile interval between 12 and 24 hours is performed to ensure consistency. This change aims to reduce API load without altering the core logic of the reconciliation loop, ensuring minimal impact on existing configurations. Developed by SIG Cloud Provider, this feature can be enabled via a specific feature gate flag. It represents a significant optimization for cloud-native infrastructure management, enhancing efficiency and reducing operational overhead for users managing Kubernetes clusters on various cloud platforms.
Kubernetes Blog