Setting Up Grafana and Prometheus Monitoring for DBLE JVM Metrics Using Docker

1 Introduction

This article demonstrates how to build a monitoring system for DBLE JVM-related metrics (CPU, Memory Pool, GC, Thread) using Docker, Grafana, and Prometheus.

Preparation

Linux server: 10.186.63.8

Docker already installed

2 Environment Setup

1. Deploy DBLE

The following steps are based on the Docker quick‑start guide.

Prepare MySQL containers

# Create a Docker network dble-net: 172.18.0.0/16
docker network create -o "com.docker.network.bridge.name"="dble-net" --subnet 172.18.0.0/16 dble-net

# Create two MySQL containers mapped to host ports 33061 and 33062
docker run --name backend-mysql1 --ip 172.18.0.2 -e MYSQL_ROOT_PASSWORD=123456 -p 33061:3306 --network=dble-net -d mysql:5.7 --server-id=1
docker run --name backend-mysql2 --ip 172.18.0.3 -e MYSQL_ROOT_PASSWORD=123456 -p 33062:3306 --network=dble-net -d mysql:5.7 --server-id=2

If the docker run... command returns a 408 Request Time‑out error, retry a few times.

Deploy DBLE

Deploy DBLE service with Docker.

docker run -d -i -t --name dble-server --ip 172.18.0.4 -p 8066:8066 -p 9066:9066 -p 8099:8099 \
    -v /opt/jmx-exporter:/jmx-exporter \
    --network=dble-net \
    actiontech/dble:latest

Port 8099 is used to expose monitoring metrics.

JMX Exporter runs as a Java agent inside the JVM, exposing metrics in Prometheus format.

# Create directory for JMX Exporter
mkdir /opt/jmx-exporter
# Download the exporter jar
wget https://repo1.maven.org/maven2/io/prometheus/jmx/jmx_prometheus_javaagent/0.15.0/jmx_prometheus_javaagent-0.15.0.jar

DBLE currently supports up to jmx_prometheus_javaagent-0.15.0.jar ; using a newer version will cause DBLE startup failure.

Add a prometheus-jmx-config.yaml file with the following content:

ssl: false
lowercaseOutputName: false
lowercaseOutputLabelNames: false

Enter the dble‑server container and modify the bootstrap configuration:

docker exec -it dble-server bash

vi /opt/dble/conf/bootstrap.cnf
# Add Java agent configuration
-javaagent:/jmx-exporter/jmx_prometheus_javaagent-0.15.0.jar=8099:/jmx-exporter/prometheus-jmx-config.yaml

Restart DBLE and verify it starts correctly.

/opt/dble/bin/dble restart

vi /opt/dble/logs/wrraper.log

View DBLE JVM metrics

Access http://10.186.63.8:8099 to see the exported metrics.

2. Deploy Prometheus

Run Prometheus in Docker:

docker run -itd --name prometheus --ip 172.18.0.5 -p 9090:9090 --network=dble-net prom/prometheus

Enter the container, edit prometheus.yml , and add a job named DBLE_Job targeting 172.18.0.4:8099 :

docker exec -it prometheus sh

vi /etc/prometheus/prometheus.yml
# Overwrite with the following configuration
scrape_configs:
  - job_name: 'DBLE_Job'
    scrape_interval: 30s
    static_configs:
    - targets:
      - '172.18.0.4:8099'

Note: 172.18.0.4 is the IP of the dble‑server container.

Restart the Prometheus container and verify the target at http://10.186.63.8:9090/targets is up.

3. Deploy Grafana

Run Grafana in Docker:

docker run -itd --name=grafana --ip 172.18.0.6 -p 3000:3000 --network=dble-net grafana/grafana

Open http://10.186.63.8:3000 and log in with the default credentials admin/admin .

4 Monitoring Dashboards

Import JVM Monitoring Template

In Grafana, add a Prometheus data source pointing to http://172.18.0.5:9090 .

Import dashboard ID 8878 (or the corresponding JSON file; another JVM template ID is 8563 ).

Customize the dashboard name as needed. The resulting view displays CPU, Memory Pool, GC, Thread and other JVM metrics.

5 Version Notes

For DBLE versions ≤ 2.20.04.0, add the Java agent configuration in the bin/wrapper.cnf file.

DBLE versions 3.20.07.0 and 3.20.10.0 do not support the -javaagent option, so JVM monitoring cannot be configured.

DBLE versions ≥ 3.21.02.0 follow the configuration described above.

wrapper.java.additional.
=-javaagent:/jmx-exporter/jmx_prometheus_javaagent-0.15.0.jar=8099:/jmx-exporter/prometheus-jmx-config.yaml

References

[1] Docker Quick Start: https://actiontech.github.io/dble-docs-cn/0.overview/0.3_quick_start/0.3.1_dble_quick_start_docker.html