Docker Monitoring with cAdvisor, Prometheus and Grafana

Docker is the open-source tool, which works like a virtual machine. However, it doesn’t create whole operating system, it has just minimum set of operating system for running application. Docker allows you to get more apps running on the server and also makes it very easy to package and ship programs.

While we have more running application, it is very important to monitor these containers performances, resource usage, etc. A container monitoring system collects metrics to ensure application running on containers are performing properly. Metrics are tracked and analyzed in real time to determine if an application is meeting expected goals.

In this article, we will cover docker monitoring by using cAdvisor, Prometheus and Grafana. These three tools are open source and easy to setup. Moreover, they can be integrated smoothly. We will setup grafana and cAdvisor using Docker container. Prometheus will run on the system.

Step-1 Install cAdvisor

cAdvisor (Container Advisor) collects, aggregates, processes and exports information about running containers. For installing cAdvisor with docker;

sudo docker run \
  --volume=/:/rootfs:ro \
  --volume=/var/run:/var/run:ro \
  --volume=/sys:/sys:ro \
  --volume=/var/lib/docker/:/var/lib/docker:ro \
  --volume=/dev/disk/:/dev/disk:ro \
  --publish=8080:8080 \
  --detach=true \
  --name=cadvisor \
  gcr.io/google-containers/cadvisor:latest 

cAdvisor is now running on http://localhost:8080

Usage Overview

CPU Usage of Containers

Step-2 Install Prometheus

Prometheus is an open-source system monitoring and alerting toolkit. Prometheus is very easy to setup. Firstly, we need to download the latest release of prometheus from https://prometheus.io/download/ .

Prometheus Download Page

Extract the downloaded file and set it as working directory.

tar xvfz prometheus-*.tar.gz
cd prometheus-*

Secondly, we need to configure prometheus.ymlbefore starting prometheus.

# global config
global:
  scrape_interval:     15s
  evaluation_interval: 15s

rule_files:
  # - "first_rules.yml"
  # - "second_rules.yml"
# A scrape configuration containing exactly one endpoint to scrape:
# Here it's Prometheus itself.
scrape_configs:
  - job_name: 'prometheus'
    static_configs:
      - targets: ['localhost:9090']
  - job_name: 'cadvisor'
    scrape_interval: 10s 
    metrics_path: '/metrics'
    static_configs:
      - targets: ['localhost:8080']
        labels:
          group: 'cadvisor'

Now, we are ready to start prometheus

./prometheus --config.file=prometheus.yml

Prometheus will start in 30 seconds. We can access prometheus main page from http://localhost:9090/metrics

Prometheus Main Page

For checking configuration on prometheus, we should follow the path: status → Targets

cAdvisor

Configuration is done. We are ready to visualize the metrics.

Step-3 Install Grafana

Grafana is an open source metric analytics & visualization suite. It is used for visualizing time series data for infrastructure and application analytics. Grafana installation is very laborless, and we will run it using the official Docker container.

$ docker run -d -p 3000:3000 grafana/grafana

Grafana is running on port 3000 so, we will use http://localhost:3000 for going main page of grafana.

Main Page of Grafana

Firstly, we need to add data source. Click on create your first data source button. Prometheus is default time series database of grafana thus, we don’t need any data source plugin.

Data Source Page

Write your prometheus URL.

Then, we can create our dashboard after adding data source. There are different kinds of dashboards with different data sources at grafana dashboard home page https://grafana.com/grafana/dashboards. At this page, we will search docker monitoring for prometheus data source. This page has so many docker monitoring dashboards, but in this tutorial, we will use docker monitoring by philicious https://grafana.com/grafana/dashboards/193.

Grafana Dashboard

We will download this dashboard as json format.

Grafana Home Page

At the main page of grafana; if we click on Home at left-top of the page, import dashboard option will become visible. Upload the downloaded json file from this page.

Import Dashboard

Name of dashboard will come automatically, but we have to select prometheus option. Then, importing is ready.

Docker Monitoring Dashboard-1

Docker Monitoring Dashboard-2

This dashboard shows us running containers, total memory usage, total cpu usage etc. As we can see there are 3 running containers which are cAdvisor, jenkins and grafana. If we want to evaluate our docker container usages, cAdvisor is using the highest CPU but jenkins has the highest memory usage.

I hope this article help you to monitor your docker containers. If you think there are questionable or missing parts, you can contact me.