How We Benchmark at Lunar

Lunar's approach to latency benchmarking highlights their commitment to maintaining minimal latency in API interactions, particularly through their Lunar Proxy solution, which mediates traffic between users' applications and third-party providers. By benchmarking their system's latency footprint, Lunar aims to ensure that their proxy adds minimal delay to API response times, even when employing remedy and diagnosis plugins designed to enhance API consumption and transaction analysis. Their experiments, conducted on AWS EC2 instances using Apache AB for client-side emulation and go-httpbin for provider-side simulation, revealed that Lunar Proxy introduces a relatively small increase in response time, with variations depending on the use of plugins. The study also showed that Lunar's latency footprint remains consistent regardless of the provider's response time. Lunar emphasizes the importance of continuous benchmarking to adapt to system changes and ensure high-quality performance, with plans to explore further aspects of benchmarking such as capacity testing in future analyses.