golang-performancelisted

# Go Performance Optimization Optimize an observed constraint, not code that merely looks slow. Preserve correctness, maintainability, and operational safety while improving the metric the user actually cares about. ## Establish the Performance Contract Before editing, record: - the workload and environment; - the primary metric, such as latency, throughput, CPU, allocation rate, resident memory, or startup time; - a representative baseline and its variance; - the target or regression threshold; - correctness and resource guardrails; - whether the bottleneck is inside this Go process or in an external dependency. If there is no useful evidence yet, gather a focused profile, trace, benchmark, or production metric first. Do not claim an optimization based only on a code pattern. ## Diagnose the Dominant Cost Use the evidence to route the investigation: | Signal | Read | | --- | --- | | CPU profile dominated by application code | [references/cpu.md](references/cpu.md) | | Allocation or retained-heap profile is high | [references/memory.md](references/memory.md) | | Time is spent waiting on network, disk, database, or syscalls | [references/io-networking.md](references/io-networking.md) | | GC, scheduler, memory limit, or build profile is implicated | [references/runtime.md](references/runtime.md) | | Repeated work, hot lookups, or duplicate concurrent requests dominate | [references/caching.md](references/caching.md) | | A production change needs metric validation | [ref