Memory Usage Health Check Component Design

Memory Usage Health Check

Memory health checks detect memory leaks and excessive memory consumption before they cause out-of-memory (OOM) kills or performance degradation.

Response Component

{
  "memory": {
    "status": "UP",
    "duration": "1ms",
    "message": "Memory usage within limits",
    "details": {
      "heapUsed": "256MB",
      "heapTotal": "512MB",
      "rss": "580MB",
      "external": "24MB",
      "utilization": "50%"
    }
  }
}

Node.js Implementation

function checkMemory() {
  const mem = process.memoryUsage();
  const heapUsedMB = Math.round(mem.heapUsed / 1024 / 1024);
  const heapTotalMB = Math.round(mem.heapTotal / 1024 / 1024);
  const utilization = (mem.heapUsed / mem.heapTotal * 100).toFixed(1);

  let status = 'UP';
  if (parseFloat(utilization) > 90) status = 'DOWN';
  else if (parseFloat(utilization) > 75) status = 'DEGRADED';

  return {
    status,
    duration: '1ms',
    message: `Heap: ${heapUsedMB}MB / ${heapTotalMB}MB (${utilization}%)`
  };
}

Memory Thresholds

Metric	Healthy	Degraded	Unhealthy
Heap utilization	< 75%	75-90%	> 90%
RSS growth rate	Stable	Slow growth	Rapid growth
GC pause time	< 50ms	50-200ms	> 200ms

Why Include in Liveness Probe

Memory checks are one of the few dependency-like checks appropriate for liveness probes. A memory leak that pushes heap usage above 90% typically requires a restart to recover, which is exactly what liveness probes are designed to trigger.

Container Memory vs App Memory

In containerized environments, check both:

• Application heap: Your runtime's managed memory
• Container limit: The cgroup memory limit set by Kubernetes

An app can be using 60% of its heap but 95% of the container limit due to native memory, thread stacks, and memory-mapped files.