NodeJS输出内存使用信息到日志

NodeJS输出内存使用信息到日志

如需转载请标明出处：http://blog.csdn.net/itas109

前言

内存信息是应用程序非常重要的信息，尤其是在排查内存相关问题的时候。
本文将介绍NodeJS如何输出内存使用信息到日志。

1. 原理简介

1.1 参数说明

process.memoryUsage()

{
  rss: 24416256,
  heapTotal: 5222400,
  heapUsed: 2987984,
  external: 1584357,
  arrayBuffers: 74935
}

1.2 实现原理

在nodejs源码中，
rss通过libuv的uv_resident_set_memory函数实现
其他的通过v8的GetHeapStatistics函数实现

static void MemoryUsage(const FunctionCallbackInfo<Value>& args) {
  Environment* env = Environment::GetCurrent(args);

  Isolate* isolate = env->isolate();
  // V8 memory usage
  HeapStatistics v8_heap_stats;
  isolate->GetHeapStatistics(&v8_heap_stats);

  NodeArrayBufferAllocator* array_buffer_allocator =
      env->isolate_data()->node_allocator();

  // Get the double array pointer from the Float64Array argument.
  Local<ArrayBuffer> ab = get_fields_array_buffer(args, 0, 5);
  double* fields = static_cast<double*>(ab->Data());

  size_t rss;
  int err = uv_resident_set_memory(&rss);
  if (err)
    return env->ThrowUVException(err, "uv_resident_set_memory");

  fields[0] = static_cast<double>(rss);
  fields[1] = static_cast<double>(v8_heap_stats.total_heap_size());
  fields[2] = static_cast<double>(v8_heap_stats.used_heap_size());
  fields[3] = static_cast<double>(v8_heap_stats.external_memory());
  fields[4] =
      array_buffer_allocator == nullptr
          ? 0
          : static_cast<double>(array_buffer_allocator->total_mem_usage());
}

2. 代码

dump.js

const process = require('process');
const fs = require('fs');

const FILE_NAME = `memory_${process.pid}.csv`;
const INTERVAL_MS = 1000; // 1s

function FormatBytes(bytes) {
    return (bytes / 1048576).toFixed(2);
}

function showMemory() {
    const memoryObj = process.memoryUsage();
    // time, rss(MB), heapTotal(MB), heapUsed(MB), external(MB), arrayBuffers(MB)
    let result = `${new Date().toLocaleString('chinese', { hour12: false })}, ${FormatBytes(memoryObj.rss)}, ${FormatBytes(memoryObj.heapTotal)}, ${FormatBytes(memoryObj.heapUsed)}, ${FormatBytes(memoryObj.external)}, ${FormatBytes(memoryObj.arrayBuffers)}\n`;
    // console.log(result);
    fs.writeFileSync(FILE_NAME, result, { flag: 'a+' });
}

showMemory();
setInterval(() => { showMemory(); }, INTERVAL_MS);

test.js

require('./dump');

3. 测试

node test.js

• 结果
  memory_12360.csv

2023-3-10 22:30:26, 23.14, 5.68, 2.69, 0.96, 0.02
2023-3-10 22:30:27, 24.53, 5.68, 2.72, 0.96, 0.02
2023-3-10 22:30:28, 24.55, 5.68, 2.72, 0.96, 0.02
2023-3-10 22:30:29, 24.64, 5.68, 2.73, 0.96, 0.02
2023-3-10 22:30:30, 24.70, 5.68, 2.74, 0.96, 0.02
2023-3-10 22:30:31, 24.71, 5.68, 2.74, 0.96, 0.02
...

License

License under CC BY-NC-ND 4.0: 署名-非商业使用-禁止演绎

如需转载请标明出处：http://blog.csdn.net/itas109

Reference:

1. https://nodejs.org/api/process.html#process_process_memoryusage
2. https://www.dynatrace.com/news/blog/understanding-garbage-collection-and-hunting-memory-leaks-in-node-js/