sigs.k8s.io/cluster-api-provider-azure@v1.14.3/docs/proposals/images/node-os-image-cache.plantuml (about) 1 @startuml 2 title Figure 1. Controller replaces Node OS image 3 4 [*] --> ChooseHealthyNode : given interval of time has passed 5 ChooseHealthyNode : Choose a healthy node to become the new prototype image 6 ChooseHealthyNode --> ShutdownChosenNode 7 8 ShutdownChosenNode : Cleanly shut down the node by cordoning and draining it 9 ShutdownChosenNode --> SnapshotChosenNode 10 11 SnapshotChosenNode : Create a snapshot of the node's disk (OS Image) 12 SnapshotChosenNode --> RestartChosenNode 13 14 RestartChosenNode : Restart the node so it can continue operating in the cluster as it was 15 RestartChosenNode --> CreateComputeImageGalleryImage 16 17 CreateComputeImageGalleryImage : Create a compute image gallery image from the snapshot 18 CreateComputeImageGalleryImage --> DeleteSnapshot 19 20 DeleteSnapshot : Delete the snapshot 21 DeleteSnapshot --> ConfigureAzureMachinePool 22 23 ConfigureAzureMachinePool : Configure the AzureMachinePool to use the new Compute Image Gallery Image 24 25 @enduml