Collecting and scraping metrics

To collect and scrape metrics from containers running on a system with Docker Desktop installed, you can use Prometheus and Container Advisor (cAdvisor). Prometheus is a powerful open source monitoring and alerting toolkit, while cAdvisor provides container users with an understanding of the resource usage and performance characteristics of their running containers.

In this section, we’ll provide a step-by-step guide to setting up Prometheus and cAdvisor to collect and scrape metrics from containers running on Docker Desktop.

Step 1 – running cAdvisor in a Docker container

cAdvisor is a Google-developed tool that collects, processes, and exports container metrics. Let’s take a look:

1. In the chapter folder, ch12, create a new subfolder called metrics:

   mkdir metrics

2. In this folder, create a file called docker-compose.yml and add the following snippet to it:

   version: '3.8'
   services:
     cadvisor:
       image: gcr.io/cadvisor/cadvisor:v0.45.0
       container_name: cadvisor
       restart: always
       ports:
       - 8080:8080
       volumes:
       - /:/rootfs:ro
       - /var/run:/var/run:rw
       - /sys:/sys:ro
       - /var/lib/docker/:/var/lib/docker:ro

3. Run cAdvisor in a Docker container using the following command:

   docker compose up cadvisor --detach

   Replace v0.45.0 with the latest cAdvisor version available on the cAdvisor repository.

   This command mounts the necessary directories from the host system and exposes cAdvisor’s web interface on port 8080.

Attention

A version lower than the one shown here will not run, for example, on a Mac with an M1 or M2 processor.

4. You can access the cAdvisor web interface by navigating to http://localhost:8080 in your browser.

Step 2 – setting up and running Prometheus

Next, let’s set up Prometheus using the following step-by-step instructions:

1. Create a subfolder called prometheus in the metrics folder.
2. In this new folder, create a prometheus.yml configuration file with the following contents:

   global:
     scrape_interval: 15s
   scrape_configs:
     - job_name: 'prometheus'
       static_configs:
         - targets: ['localhost:9090']
     - job_name: 'cadvisor'
       static_configs:
         - targets: ['host.docker.internal:8080']

   This configuration specifies the global scrape interval and two scrape jobs: one for Prometheus itself and another for cAdvisor running on port 8080.

3. Add the following snippet to the end of the docker-compose.yml file:

   prometheus:
     image: prom/prometheus:latest
     container_name: prometheus
     restart: always
     ports:
       - 9090:9090
     volumes:
       - ./prometheus:/etc/prometheus
       - prometheus_data:/prometheus

   This instruction mounts the prometheus.yml configuration file and exposes Prometheus on port 9090.

4. The preceding prometheus service uses a volume called prometheus_data. To define this, please add the following two lines to the end of the docker-compose.yml file:

   volumes:
     prometheus_data:

5. You can access the Prometheus web interface by navigating to http://localhost:9090 in your browser.

Once Prometheus is up and running, you can verify that it’s successfully scraping metrics from cAdvisor:

1. Open the Prometheus web interface at http://localhost:9090.
2. Click on Status in the top navigation bar, then select Targets.
3. Ensure that both the prometheus and cadvisor targets are listed with State set to UP.

Now, Prometheus can collect and store metrics from the containers running on your Docker Desktop system. You can use Prometheus’ built-in expression browser to query metrics or set up Grafana for advanced visualization and dashboarding:

1. In the query text field, enter something like container_start_time_seconds to get the value for the startup time of all containers.
2. To refine the query and only get the value for the cAdvisor container, enter container_start_time_seconds{job="cadvisor"}.

Note that in the query text field, you get IntelliSense, which is convenient when you do not remember all the details of a command and its parameters.

Before you continue, stop cAdvisor and Prometheus with the following command:

docker compose down -v

In the last section of this chapter, you will learn how to monitor a containerized application using a tool such as Grafana.