Multi-Cloud Cost Optimization Pipeline with Dynamic Rightsizing

Design an advanced cost optimization automation script that continuously monitors AWS, GCP, and Azure compute resources, analyzing utilization metrics, predicting workload patterns, and automatically recommending or executing instance resizing. The solution must handle complex scenarios like microservice architectures, handle multi-region deployments, generate cost-saving reports, and integrate with existing infrastructure-as-code frameworks like Terraform. Include machine learning prediction models for forecasting future resource requirements and potential cost anomalies.