Docs

Docs

Prometheus Integration

Learn how to export Observability Kit metrics to Prometheus from a Vaadin application.

Prometheus is a pull-based metrics system: it scrapes a metrics endpoint that your application exposes. Micrometer has a native Prometheus registry, so exposing the kit’s metrics is a matter of adding the registry and an Actuator endpoint.

Prometheus handles metrics only. To collect traces as well, pair it with a tracing backend — see Jaeger & Prometheus for a local setup, or Grafana for metrics, traces, and logs in one tool.

Expose the Metrics Endpoint

Add Spring Boot Actuator and the Prometheus registry:

Source code
pom.xml
<dependency>
    <groupId>org.springframework.boot</groupId>
    <artifactId>spring-boot-starter-actuator</artifactId>
</dependency>
<dependency>
    <groupId>io.micrometer</groupId>
    <artifactId>micrometer-registry-prometheus</artifactId>
</dependency>

Then expose the Prometheus endpoint:

Source code
application.properties
management.endpoints.web.exposure.include=prometheus
spring.application.name=vaadin

The metrics are now served at GET /actuator/prometheus. Setting spring.application.name tags every meter with a consistent service name.

Start your application, exercise a few views, then open http://localhost:8080/actuator/prometheus and confirm the Vaadin meters appear — for example vaadin_sessions_active or vaadin_request_duration_seconds_count.

Scrape the Endpoint

Configure Prometheus to scrape the Actuator endpoint. The scrape path is /actuator/prometheus, not the Prometheus default of /metrics:

Source code
prometheus.yml
global:
  scrape_interval: 15s # Default is every 1 minute.

scrape_configs:
  - job_name: 'vaadin'
    metrics_path: '/actuator/prometheus'
    static_configs:
      - targets: ['localhost:8080']

If your application uses Spring Security, the scrape also needs credentials — see Secure the Endpoint.

Download Prometheus

Extract the archive, then start Prometheus with the configuration file:

Source code
terminal
./prometheus --config.file=/PATH/TO/prometheus.yml

Remember to substitute the correct path to the prometheus.yml file.

Secure the Endpoint

In an application that uses Spring Security, the Vaadin security configuration also handles /actuator/prometheus, and redirects an unauthenticated request to the login view. You can see this with curl:

Source code
terminal
curl -i http://localhost:8080/actuator/prometheus

A 302 response with a Location header pointing at the login view means the scrape never reaches the endpoint. Prometheus follows the redirect, receives the login page instead of metrics, and fails to parse it. No Vaadin series appear, and the only trace of the problem is the scrape error on the Targets page of Prometheus.

First, give the Actuator endpoints a security filter chain of their own with HTTP Basic authentication, as shown in Secure the Actuator Endpoints. Check the chain with curl -u USER:PASSWORD http://localhost:8080/actuator/prometheus, which should now return the metrics, while the same request without -u returns 401.

Then add the credentials of the scraper account to the scrape job:

Source code
prometheus.yml
scrape_configs:
  - job_name: 'vaadin'
    metrics_path: '/actuator/prometheus'
    basic_auth:
      username: 'prometheus'
      password_file: '/PATH/TO/password.txt'
    static_configs:
      - targets: ['localhost:8080']

The password file holds the account’s password and nothing else. Restart Prometheus, or reload its configuration, and check that the target shows as UP on the Targets page.

Viewing Metrics

The default address for the Prometheus user interface is http://localhost:9090. Enter a metric name into the search field and click Execute.

If you don’t know the metric name, click the globe icon next to Execute to open the metrics explorer, which lists the available metrics alphabetically.

To view a result as a graph, click the Graph tab.

Prometheus renders Micrometer’s dotted meter names in its own convention. For example, the vaadin.request.duration timer appears as vaadin_request_duration_seconds with _count, _sum, and _max suffixes — plus a parallel vaadin_request_duration_active_seconds long-task family when tracing is on, and _bucket series once you opt into histogram buckets. See the Reference page for the full list of meters.