Method AddLatency

cpp/src/arrow/flight/flight_benchmark.cc:111–114

Invoked per batch in the test loop. Holding a lock looks not scalable. Tested with 1 ~ 8 threads, no noticeable overhead is observed. A better approach may be calculate per-thread quantiles and merge.

  // Tested with 1 ~ 8 threads, no noticeable overhead is observed.
  // A better approach may be calculate per-thread quantiles and merge.
  void AddLatency(uint64_t elapsed_nanos) {
    std::lock_guard<std::mutex> lock(this->mutex);
    latencies.Add(static_cast<double>(elapsed_nanos));
  }

  // ns -> us
  uint64_t max_latency() const { return latencies.Max() / 1000; }