You are managing a critical service running on Google Kubernetes Engine (GKE) that requires consistent performance. The service is sensitive to CPU and memory fluctuations. Which of the following strategies should you implement to maintain the expected performance during GKE's automatic upgrades?