Optimizing latency

A retail application experiences a traffic spike during a flash sale. Each product page request follows the same critical path: the application validates the request, retrieves product catalog data from a database, calls an external service for additional details, and returns the response. Under load, response times increase even though compute resources are scaling.

Based on the concept of the critical path, which optimization would most effectively reduce user-facing latency?