Mastering Micrometer: From Counters to Grafana Dashboards in Spring Boot

JVM Metrics Framework Micrometer in Practice

Spring Actuator collects metrics and exposes them via Prometheus and Grafana for monitoring production environments. Micrometer provides various Meter types such as Counter, Timer, Gauge, DistributionSummary, LongTaskTimer, FunctionCounter, FunctionTimer, and TimeGauge.

MeterRegistry

MeterRegistry is an abstract class with implementations like SimpleMeterRegistry, CompositeMeterRegistry, and a global CompositeMeterRegistry. SimpleMeterRegistry stores metrics in memory, while CompositeMeterRegistry aggregates multiple registries.

MeterRegistry registry = new SimpleMeterRegistry();
Counter counter = registry.counter("counter");
counter.increment();

Tag Naming and NamingConvention

Meter names use dot notation and can be transformed for different monitoring systems via NamingConvention. Common conventions include dot, snake_case, camelCase, upperCamelCase, and slashes. Global tags can be added to all meters for dimensions like stack, environment, or region.

registry.config().commonTags("stack", "prod", "region", "us-east-1");

Meter Types and Usage

Counter

Counter records a monotonically increasing count, suitable for events like order creation or HTTP request totals.

Metrics.counter("order.create", "channel", order.getChannel(), "createTime", formatter.format(order.getCreateTime())).increment();

FunctionCounter

FunctionCounter derives its value from a function, allowing external state to drive the count.

FunctionCounter.builder("functionCounter", atomicInt, AtomicInteger::get)
    .description("functionCounter")
    .register(new SimpleMeterRegistry());
atomicInt.incrementAndGet();

Timer

Timer measures the duration of short-lived events. It can be used directly or via functional wrappers.

Timer timer = Metrics.timer("timer", "createOrder", "cost");
timer.record(() -> createOrder(order));

LongTaskTimer

LongTaskTimer records the duration of long-running tasks until they complete.

LongTaskTimer longTaskTimer = registry.more().longTaskTimer("longTaskTimer");
longTaskTimer.record(() -> {/* task logic */});

Gauge

Gauge provides a snapshot of a value that can go up and down, such as memory usage or collection sizes.

Gauge gauge = registry.gauge("queue.size", queue, Collection::size);
AtomicInteger n = registry.gauge("numberGauge", new AtomicInteger(0));

DistributionSummary

DistributionSummary tracks the distribution of non‑time values, like payload sizes.

DistributionSummary summary = DistributionSummary.builder("response.size")
    .baseUnit("bytes")
    .register(registry);
summary.record(payloadSize);

Integrating Micrometer with Spring Boot, Prometheus, and Grafana

Spring Boot's

spring-boot-starter-actuator

includes Micrometer support. Adding

micrometer-registry-prometheus

exposes a

/management/prometheus

endpoint that Prometheus scrapes.

management:
  endpoints:
    web:
      exposure:
        include: '*'
  endpoint:
    prometheus:
      enabled: true

Prometheus configuration (prometheus.yml) defines a job that scrapes the Spring Boot application's Prometheus endpoint.

scrape_configs:
  - job_name: 'spring-app'
    metrics_path: /management/prometheus
    static_configs:
      - targets: ['localhost:10091']

Grafana is added as a data source pointing to Prometheus (default port 9090). Dashboards can query metrics such as

order_count_total

or

method_cost_time_seconds_sum

to visualize order counts and method execution times.

SELECT sum(rate(order_count_total[1m])) FROM metric

By combining Micrometer, Spring Boot Actuator, Prometheus, and Grafana, developers gain fine‑grained, multi‑dimensional observability of backend services.