> ## Documentation Index > Fetch the complete documentation index at: https://docs.healops.ai/llms.txt > Use this file to discover all available pages before exploring further. # Right Sizing > Optimize resource allocation based on actual usage patterns Right-Sizing in Tracer optimizes scientific compute by using actual CPU and RAM usage from previous runs rather than static rules or guesswork. After each workflow execution, Tracer analyzes per-step utilization and compares it with current cloud instance pricing to recommend the most cost-efficient instance for the next run. Rightsizing Tracer uses statistical models to predict RAM requirements and account for variability, minimizing memory-related failures that AWS’s native tools cannot anticipate. The system detects idle CPU, excess memory headroom, and under-provisioned configurations, then maps each workflow to an instance type that fits both performance and cost. As more runs complete, recommendations become increasingly accurate, lowering compute spend, reducing over-provisioning, and preventing failures caused by mis-sized infrastructure.