Explain Kubernetes cluster autoscaling and how it integrates with cloud provider APIs.

Cluster Autoscaler automatically adjusts cluster size by adding or removing nodes based on pod scheduling requirements and resource utilization.\n\nHow Cluster Autoscaler Works:\n• Monitors unschedulable pods due to resource constraints\n• Evaluates node groups for scaling up\n• Removes underutilized nodes after scale-down delay\n• Integrates with cloud provider APIs\n• Respects node group min/max constraints\n\nScaling Decisions:\nScale Up Triggers:\n• Pods in Pending state due to insufficient resources\n• Resource requests cannot be satisfied\n• No suitable nodes available\n\nScale Down Triggers:\n• Node utilization below threshold (default 50%)\n• All pods can be moved to other nodes\n• Node has been underutilized for scale-down delay\n\nCloud Provider Integration:\nAWS:\n• Auto Scaling Groups (ASG)\n• EC2 instance management\n• Spot instance support\n• Mixed instance types\n\nGCP:\n• Managed Instance Groups (MIG)\n• Preemptible instance support\n• Regional persistent disks\n\nAzure:\n• Virtual Machine Scale Sets (VMSS)\n• Spot instance integration\n• Availability zones support\n\nBest Practices:\n• Set appropriate resource requests\n• Use pod disruption budgets\n• Monitor scaling events and costs\n• Configure node group diversity\n• Implement proper tagging strategies\n• Use Vertical Pod Autoscaler alongside