stevenacoffman/prometheus-spring-actuator repository overview

⁠Example Spring Boot Actuator app with Prometheus on Kubernetes

inspired by

• https://reflectoring.io/monitoring-spring-boot-with-prometheus/⁠
• https://github.com/prometheus/client_java⁠
• https://github.com/microservices-demo/carts⁠
• http://blog.monkey.codes/actuator-and-prometheus/⁠
• https://github.com/f-lopes/spring-boot-docker⁠
• https://github.com/spotify/dockerfile-maven⁠

Do note that this does not seem to work well with Spring Boot 2.x (e.g. 2.0.0.M6)

If you wish to run Grafana and Prometheus (https://prometheus.io/docs/visualization/grafana/⁠) locally, you'll need to install them.

To run Grafana locally, use:

grafana-server --config=/usr/local/etc/grafana/grafana.ini --homepath /usr/local/share/grafana cfg:default.paths.logs=/usr/local/var/log/grafana cfg:default.paths.data=/usr/local/var/lib/grafana cfg:default.paths.plugins=/usr/local/var/lib/grafana/plugins

And then visit http://0.0.0.0:3000/login⁠ with admin / admin

Once this application is running, visit http://localhost:8080/health⁠

Custom metrics we're interested in having in the near future

• gc
• other JMX stats?
• error rate
• transaction rate (MVC calls, ...)

⁠Lightweight Spring Boot Docker image based on Alpine + Docker Compose file

Features:

• Spring profiles
• Custom JAVA_OPTS
• Healthcheck
• Wrapper script that sets (or not) debug and jmx java options
• Debug mode

⁠Available environment variables

⁠How to use ?

1. Build the image with mvn package and push it with mvn deploy. Of course you can also say mvn dockerfile:build explicitly. This uses dockerfile-maven⁠.

2. Start your application:

   • Using Docker CLI

   • Using the provided docker-compose.yml in this repository:

docker-compose -f compose/docker-compose.yml up

⁠Inject environment variables:

docker run -d -p 8080:8080 -e JAVA_OPTS=-Xms256m -Xmx512m spring-boot-image

⁠Using Docker Compose

The docker-compose.yml file picks its configuration from the .env one.

Using this file, you can set the desired properties:

# Docker
IMAGE_NAME=stevenacoffman/prometheus-spring-actuator


# Application
SERVER_PROTOCOL=http
SERVER_PORT=8080
DEBUG_PORT=8000
JMX_PORT=9010
SPRING_PROFILES_ACTIVE=dev
DEBUG=false
JMX=false
# -Dprop=... -Dparameter=...
JAVA_OPTS=

# Host
HOST_SERVER_PORT=8080
HOST_DEBUG_PORT=8000
HOST_JMX_PORT=9010

⁠Debug mode

To run the application in debug mode, simply set the DEBUG variable to true in the .env file.

Launch the application: docker-compose -f compose/docker-compose.yml up

By default, the debug port used is the 8000.

⁠JMX mode

To run the application in JMX mode, simply set the JMX variable to true in the .env file.

Launch the application: docker-compose -f compose/docker-compose.yml up

By default, the debug port used is the 9010.

⁠Spring profile(s)

By default, the application will run with dev Spring profile

To run the application with a specific Spring profile, set the desired one in the .env file:

SPRING_PROFILES_ACTIVE=dev

⁠Running the application via Kubernetes

$ kubectl apply -f ./k8s

⁠Making Prometheus In Kubernetes Automatically pick up endpoint

Add the proper annotations to your deployment to inform prometheus what and where to scrape:

spec:
  replicas: 1 # tells deployment to run 2 pods matching the template
  template: # create pods using pod definition in this template
    metadata:
      labels:
        run: prometheus-spring-actuator
      annotations:
        prometheus.io/scrape: "true"
        prometheus.io/port: "8888"