Skip to main content
A complete walkthrough of instrumenting a Java service using the official Prometheus Java client library (client_java v1.x). See the overview for the general approach.

​
1. Add the Prometheus client dependency

Add these dependencies to your pom.xml (Maven): 
<dependency>
    <groupId>io.prometheus</groupId>
    <artifactId>prometheus-metrics-core</artifactId>
    <version>1.3.6</version>
</dependency>
<dependency>
    <groupId>io.prometheus</groupId>
    <artifactId>prometheus-metrics-instrumentation-jvm</artifactId>
    <version>1.3.6</version>
</dependency>
<dependency>
    <groupId>io.prometheus</groupId>
    <artifactId>prometheus-metrics-exporter-httpserver</artifactId>
    <version>1.3.6</version>
</dependency>
Or with Gradle (build.gradle): 
implementation 'io.prometheus:prometheus-metrics-core:1.3.6'
implementation 'io.prometheus:prometheus-metrics-instrumentation-jvm:1.3.6'
implementation 'io.prometheus:prometheus-metrics-exporter-httpserver:1.3.6'

​
2. Register custom metrics and expose /metrics

Register default JVM metrics, define your own custom counters/gauges, and start an HTTP server that exposes them on /metrics: 
import io.prometheus.metrics.core.metrics.Counter;
import io.prometheus.metrics.exporter.httpserver.HTTPServer;
import io.prometheus.metrics.instrumentation.jvm.JvmMetrics;

public class App {

    static final Counter emailsSent = Counter.builder()
        .name("emails_sent_total")
        .help("Total number of emails sent")
        .labelNames("status")
        .register();

    public static void main(String[] args) throws Exception {
        // Register out-of-the-box JVM metrics (memory, GC, threads, classloader, etc.)
        JvmMetrics.builder().register();

        // Start the metrics HTTP server on port 9090
        HTTPServer.builder()
            .port(9090)
            .buildAndStart();

        // Increment your custom metric anywhere in your code
        emailsSent.labelValues("success").inc();
    }
}
This starts a separate HTTP server on port 9090 that responds to GET /metrics in Prometheus exposition format — including JVM metrics like jvm_memory_used_bytes, jvm_gc_pause_seconds, jvm_threads_live_threads, plus your custom emails_sent_total.
If you’re using Spring Boot, the typical approach is Spring Boot Actuator + Micrometer’s Prometheus registry. Add io.micrometer:micrometer-registry-prometheus and set management.endpoints.web.exposure.include=prometheus in application.properties. The /actuator/prometheus endpoint is then exposed on your main service port.

​
3. Declare the metrics endpoint in ops.json

{
    "metrics": {
        "endpoint": "/metrics",
        "interval": 15,
        "port": 9090
    }
}
Prometheus will now scrape http://<your-service>:9090/metrics every 15 seconds and make every metric — JVM and custom — available in Grafana.

​
4. Visualize JVM metrics with a community dashboard

Instead of building JVM dashboards from scratch, import a battle-tested community dashboard from grafana.com/grafana/dashboards.
DashboardIDNotes
JVM (Micrometer)4701Most popular JVM dashboard. Built for Micrometer but most panels work with client_java v1.x since metric names like jvm_memory_used_bytes and jvm_threads_live_threads overlap.
JVM (Micrometer) — alternative8898Alternative JVM dashboard, also Micrometer-based.
JVM SpringBoot322108For Spring Boot 3.x apps using Actuator + Micrometer.
To import: in Grafana, go to Dashboards → New → Import, enter the dashboard ID, and select your LocalOps Prometheus data source. You’ll instantly see heap/non-heap memory, GC activity, thread counts, classloader stats, and more. Your custom application metrics (emails_sent_total, etc.) can be charted alongside in the same or a separate dashboard.