You want to reason about the problem and test your hypothesis and trade-offs.
- Identify: Use benchmarking tests identify and measure the performance regression or runtime pprof tools to create a callgraph of the executable with the given dataset that is problematic.
- Hypotheses: Based on the benchmarks or callgraph, find out what you think is the cause of the regression.
- Test: Isolate (if possible) the problematic code, create a minimal
reproducible example, and test using
- Evaluate and share: Record your process, findings, and share results.
This can be translated into practical steps using go tools and related libraries.