Benchmark Report —

A benchmark without a reproducible harness is not accepted. Link the harness.

Methodology

  • What is measured: (throughput / latency / memory / cold start)
  • Harness: link to the script/repo (must be open and re-runnable)
  • Load generator & settings: (tool, concurrency, duration, warmup)
  • What is NOT measured / caveats:

Hardware / environment

  • CPU, memory, OS:
  • Node version, StreetJS version:
  • Database/cache versions (if in scope):
  • Network (same-host / LAN / cloud region):

Workloads

Describe each workload (payload size, route shape, DB involvement).

Results (raw)

| Workload | Throughput (req/s) | p50 | p95 | p99 | memory | |———-|——————–|—–|—–|—–|——–|

Attach the raw output (or link it). Include run-to-run variance.

Comparison (optional)

If comparing to another framework, run both with the same harness and publish both configs. State versions and any tuning applied to each.

Reproduction 

1

# exact commands to reproduce