golang-performance

Profile Go performance bottlenecks using pprof or fgprof to identify whether latency stems from database wait time, excess allocations, or bad data layout, then apply one targeted optimization and verify gains with benchstat.

Reducing allocations on a Go hot path requires profiling to confirm the bottleneck, then applying proven patterns like slice reuse, avoiding interface boxing, and improving cache locality, validated through iterative benchmarking.

Benchmark a Go optimization by measuring a baseline, applying one change at a time, and verifying gains with statistical comparison using benchstat to ensure disciplined, measurable performance improvement.

Tune the Go garbage collector when profiling with pprof indicates GC pressure is the primary bottleneck, applying runtime tuning and concurrency-aware optimization patterns to improve latency and CPU efficiency.

Profiling-driven optimization applies to production Go workloads where latency, memory, CPU, GC, or I/O behavior must improve, using disciplined measurement with pprof, tracing, and benchmark-driven analysis.

Diagnose a slow Go service under load by profiling to determine the real bottleneck, checking for database wait time, excess allocations, inefficient loops, or bad data layout before changing code.