Micrometer

Language: Java

Monitoring / Metrics

Micrometer was developed to unify metrics collection in Java applications, particularly for modern microservices architectures. It integrates seamlessly with Spring Boot and provides adapters for popular monitoring systems like Prometheus, Graphite, Datadog, and New Relic. This enables developers to monitor performance, track business metrics, and diagnose issues in production environments.

Micrometer is a metrics instrumentation library for Java applications that provides a vendor-neutral interface for capturing application metrics and exposing them to monitoring systems. It supports counters, gauges, timers, distribution summaries, and more.

Installation

maven:

Add dependency in pom.xml:
<dependency>
  <groupId>io.micrometer</groupId>
  <artifactId>micrometer-core</artifactId>
  <version>1.12.1</version>
</dependency>

gradle: implementation 'io.micrometer:micrometer-core:1.12.1'

Usage

Micrometer provides a consistent API to record metrics like counters, gauges, timers, and distribution summaries. Metrics can be tagged for filtering and organized for export to external monitoring systems.

Creating a counter

import io.micrometer.core.instrument.MeterRegistry;
import io.micrometer.core.instrument.Counter;

Counter counter = registry.counter("requests_total");
counter.increment();

Creates a counter metric and increments it to track the total number of requests.

Creating a gauge

import io.micrometer.core.instrument.Gauge;
import java.util.concurrent.atomic.AtomicInteger;

AtomicInteger queueSize = new AtomicInteger(0);
Gauge.builder("queue.size", queueSize, AtomicInteger::get)
     .register(registry);

Creates a gauge metric to monitor the current size of a queue.

Using a timer

import io.micrometer.core.instrument.Timer;

Timer timer = registry.timer("requests_latency");
timer.record(() -> {
    // code to measure
});

Measures the execution time of a code block and records it as a timer metric.

Distribution summary

import io.micrometer.core.instrument.DistributionSummary;

DistributionSummary summary = DistributionSummary.builder("payload.size").register(registry);
summary.record(1024);

Records numeric values to track distributions, e.g., payload sizes.

Tagging metrics

Counter counter = Counter.builder("requests_total")
    .tags("endpoint", "/api/users", "status", "success")
    .register(registry);
counter.increment();

Adds tags to metrics for filtering and aggregation in monitoring systems.

Publishing to Prometheus

import io.micrometer.prometheus.PrometheusMeterRegistry;

PrometheusMeterRegistry prometheusRegistry = new PrometheusMeterRegistry(PrometheusConfig.DEFAULT);

Configures Micrometer to expose metrics in Prometheus format for scraping by monitoring tools.

Error Handling

MeterAlreadyExistsException: Occurs if a metric with the same name and tags is registered multiple times. Reuse existing metrics or use unique tags.

No registry configured: Ensure a MeterRegistry instance is available and properly configured before registering metrics.

Best Practices

Instrument all important business and technical metrics using counters, gauges, and timers.

Use tags to differentiate metrics by dimensions like endpoint, region, or user.

Export metrics to a centralized monitoring system for alerting and visualization.

Avoid high-cardinality tags that can overwhelm the monitoring backend.

Leverage Micrometer integration with Spring Boot Actuator for automatic metrics collection.

Official Docs Github