Metrics with Micrometer in Spring Boot - Performance & Optimization

Performance: Metrics with Micrometer

MEDIUM IMPACT

This concept affects how efficiently application metrics are collected and reported, impacting monitoring overhead and response time.

Collecting application metrics for monitoring

Spring Boot

MeterRegistry registry = new PrometheusMeterRegistry(PrometheusConfig.DEFAULT);
Counter counter = registry.counter("requests");
for (int i = 0; i < 10000; i++) {
  counter.increment();
}
// Metrics are scraped asynchronously by Prometheus

Using a Prometheus registry enables asynchronous scraping, reducing blocking and CPU overhead during metric updates.

📈 Performance GainNon-blocking metric collection, reducing CPU spikes and improving app responsiveness

Collecting application metrics for monitoring

Spring Boot

MeterRegistry registry = new SimpleMeterRegistry();
Counter counter = registry.counter("requests");
for (int i = 0; i < 10000; i++) {
  counter.increment();
}

Using a simple in-memory registry without asynchronous reporting causes high CPU usage and blocks threads during metric updates.

📉 Performance CostBlocks application threads during increments, increasing response latency under load

Performance Comparison

Pattern	CPU Usage	Memory Usage	Blocking	Verdict
SimpleMeterRegistry with synchronous increments	High CPU during increments	Low memory	Blocks threads on increment	[X] Bad
PrometheusMeterRegistry with async scraping	Low CPU spikes	Moderate memory for buffering	Non-blocking increments	[OK] Good

Rendering Pipeline

Micrometer metrics collection runs alongside application code, affecting CPU and memory usage but not browser rendering directly.

→Application CPU

→Memory Usage

⚠️ BottleneckSynchronous metric updates causing thread blocking

Optimization Tips

1Avoid synchronous metric updates on hot code paths to prevent blocking.

2Use asynchronous or batch reporting registries like PrometheusMeterRegistry.

3Profile CPU and thread usage to detect metric collection overhead.

Performance Quiz - 3 Questions

What is a key performance benefit of using PrometheusMeterRegistry with Micrometer?

A It stores all metrics in memory permanently

B It allows asynchronous metric scraping, reducing blocking in the app

C It disables metric collection to save CPU

D It increases CPU usage to improve accuracy

DevTools: JVisualVM or Java Flight Recorder

How to check: Run your Spring Boot app with Micrometer, profile CPU and thread activity during metric increments, and observe blocking or CPU spikes.

What to look for: Look for thread blocking during metric updates and CPU usage spikes indicating synchronous metric overhead.