I would install benchmarking tools across every aspect of the code pipeline starting with the developers IDE. Performance optimization checks on the local dev environment, at Dev and QA Tiers as well as Performance environment. In addition, I would define performance gateways driven by benchmarks scores that prevent code from being elevated beyond QA tier if it does not meet performance criteria. These kinds of gateways can be defined very discretely in tools like Sonarqube. This FOSS platform will scan your entire code base and call out hundreds of performance violation. Enterprises can reject code based on the scan results. The platform goes as far as explaining, with sample code, how the code should be re-written to be more performant. Once the code gets to an integrated Tier env, I would run integration related performance testing and trigger E2E. Thinking I crushed it, I paused, he seemed annoyed as he wanted me to mention one specific tool he had in mind and cared nothing about my holistic answer. The entire time I was speaking he nodded his head repeatedly and kept interrupting with non-verbals even though he did not ask a question.