Skip to main content

System Metrics

Monitor your server's health with system-level metrics.

Available Metrics​

CPU​

  • CPU Usage % - Percentage of CPU utilized
  • CPU User - User-space CPU time
  • CPU System - Kernel CPU time

Memory​

  • Memory Used - RAM currently in use
  • Memory Free - Available RAM
  • Heap Used - Node.js heap usage
  • Heap Total - Total heap allocated

Event Loop​

  • Event Loop Lag - Delay in the Node.js event loop
  • Active Handles - Open handles (sockets, timers, etc.)
  • Active Requests - Pending async operations

MongoDB​

  • Connection Pool Size - Current/max connections
  • Available Connections - Idle connections
  • Query Time - Database query latency

Viewing System Metrics​

System Tab​

Navigate to your site's System tab to see:

  • Real-time metric charts
  • Per-host breakdown
  • Historical trends

Host Comparison​

For multi-server deployments:

  1. Select "All Hosts" in the host filter
  2. Enable "Compare Hosts" toggle
  3. See metrics side-by-side

Metric Details​

CPU Usage​

High CPU can indicate:

  • Heavy computation in methods
  • Inefficient algorithms
  • Too many concurrent operations
// Bad: CPU-heavy operation in method
Meteor.methods({
  'data.process': function(data) {
    // This blocks the event loop
    return heavyComputation(data);
  }
});

// Better: Offload to background job
Meteor.methods({
  'data.process': function(data) {
    Jobs.run('processData', { data });
    return { status: 'processing' };
  }
});

Memory Usage​

Monitor for:

  • Memory leaks - Steadily increasing usage
  • High heap - Large in-memory datasets
  • GC pressure - Frequent garbage collection

Event Loop Lag​

Event loop lag > 100ms indicates issues:

  • Long-running synchronous operations
  • Heavy computation
  • Too many concurrent operations
// Bad: Blocks event loop
const result = heavySync();

// Good: Break up work
async function processInChunks(items) {
  for (const chunk of _.chunk(items, 100)) {
    await processChunk(chunk);
    await new Promise(r => setTimeout(r, 0)); // Yield
  }
}

MongoDB Metrics​

Watch for:

  • Pool exhaustion - All connections busy
  • High query time - Slow database queries
  • Connection errors - Database connectivity issues

Setting Up Alerts​

Create alerts for system metrics:

  1. Go to Site Settings > Alerts
  2. Create alert with conditions like:
    • CPU > 80% for 5 minutes
    • Memory > 90%
    • Event loop lag > 500ms
    • MongoDB connections > 80%

Best Practices​

1. Baseline Your Metrics​

Understand normal values before alerting:

  • What's typical CPU during peak hours?
  • How much memory does your app normally use?
  • What's normal event loop lag?

Look for patterns:

  • Memory slowly increasing (leak?)
  • CPU spikes at certain times (cron jobs?)
  • Connection pool filling up (query issues?)

3. Correlate with Events​

Compare system metrics with:

  • Deployments
  • Traffic changes
  • Error spikes
  • Method response times

Next Steps​