Every so often, you’ll read a performance benchmark, with bold claims for how X is better than Y. However, the process is surprisingly difficult to do well, with many opportunities for mistakes and oversights Metrics can be confusing, complicated, unreliable, inaccurate, and even plain wrong (due to bugs). Benchmarking in its nature requires a **complete** understanding of: 1) **What** you are benchmarking 2) **Why** you are benchmarking 3) What are the application **limitations**/**boundaries** 4) **How** you are benchmarking Benchmarking shouldn’t be complex but certainly isn’t a fire-and-forget activity. When dealing with runtimes such as Node.js a lot of variables should be considered when benchmarking, there is also, hidden behaviors from the runtime you should be aware. Benchmark is an art. Throughout this talk, some of the mysteries behind benchmarks will be covered and new ones, eventually, will arise.