Add documentation for AutoScaleFromZero

Add documentation for the `AutoScaleFromZero` feature flag

Author: nikParasyr

docs/book/src/SUMMARY.md

@@ -5,6 +5,7 @@
 - [Configuration](clusteropenstack/configuration.md)
 - [Experimental Features](./experimental-features/experimental-features.md)
     - [PriorityQueue](./experimental-features/priority-queue.md)
+    - [AutoScaleFromZero](./experimental-features/autoscale-from-zero.md)
 - [Topics](./topics/index.md)
     - [external cloud provider](./topics/external-cloud-provider.md)
     - [hosted control plane](./topics/hosted-control-plane.md)

docs/book/src/experimental-features/autoscale-from-zero.md

@@ -0,0 +1,39 @@
+# AutoScale From Zero
+
+> **Note:** AutoScaleFromZero is available in >= 0.14.
+
+> **Note**: AutoScaleFromZero can be used only in clusters using the [ClusterClass](https://cluster-api.sigs.k8s.io/tasks/experimental-features/cluster-class/) feature.
+
+The `AutoScaleFromZero` feature flag enables the usage of [cluster-autoscaler](https://github.com/kubernetes/autoscaler/tree/bc3f44c85df17bccc940adb7c885b192cf6135d7/cluster-autoscaler/cloudprovider/clusterapi#cluster-autoscaler-on-cluster-api) to scale from/to zero without the need of annotations. More information on how to use the cluster-autoscaler can be found [here](https://github.com/kubernetes/autoscaler/tree/bc3f44c85df17bccc940adb7c885b192cf6135d7/cluster-autoscaler/cloudprovider/clusterapi#scale-from-zero-support).
+
+## Enabling AutoScaleFromZero
+
+You can enable `AutoScaleFromZero` using the following.
+
+- Environment variable: `EXP_CAPO_AUTOSCALE_FROM_ZERO=true`
+- clusterctl.yaml variable: `EXP_CAPO_AUTOSCALE_FROM_ZERO: true`
+- --feature-gates argument: `AutoScaleFromZero=true`
+
+## Automatically Populated Status Fields
+
+> **Note**: Unsupported fields may be provided via annotations or incorporated into the controller by extending its functionality.
+
+The controller automatically fills two sections of `OpenStackMachineTemplate.Status`:  
+- **capacity** (resource quantities)  
+- **nodeInfo** (OS metadata)  
+
+The following mappings describe exactly where each value originates.
+
+### Capacity (`Status.Capacity`)
+- **CPU**: From the `VCPUs` property of the resolved OpenStack flavor
+
+- **Memory**: From the `RAM` property of the resolved OpenStack flavor
+
+- **Ephemeral Storage**: From the `Ephemeral` property of the resolved OpenStack flavor
+
+- **Root Storage**: Determined based on the boot method:  
+  - If **booting from volume** taken from `OpenStackMachineTemplate.Spec.Template.Spec.RootVolume.SizeGiB`  
+  - If **booting from image** taken from the `Disk` property of the resolved OpenStack flavor
+
+### Node Information (`Status.NodeInfo`)
+- **Operating System**: From the `os_type` property of the resolved OpenStack image.

docs/book/src/experimental-features/experimental-features.md

@@ -4,6 +4,7 @@ CAPO now ships with experimental features the users can enable.
 
 Currently CAPO has the following experimental features:
 * `PriorityQueue` (env var: `EXP_CAPO_PRIORITY_QUEUE`): [PriorityQueue](./priority-queue.md)
+* `AutoScaleFromZero` (env var: `EXP_CAPO_AUTOSCALE_FROM_ZERO`): [AutoScaleFromZero](./autoscale-from-zero.md)
 
 ## Enabling Experimental Features for Management Clusters Started with clusterctl