/docker-monitor

Scripts and configuration to help you monitor your Docker containers

Primary LanguageShell

Docker Container Monitoring

Monitor real-time and historical memory, cpu, and network usage of your Docker containers.

Run cAdvisor with InfluxDB and Grafana

Launch the cAdvisor monitoring service with an InfluxDB storage backend and Grafana web-based dashboard with a simple command. Everything runs in Docker containers, so you don't need to install anything. Several Grafana dashboards are created for you: one showing all containers, and one for each.

docker-compose.yml is a Docker Compose definition that starts the three services in containers and wires them together for you.

Read docker-compose.README.md for details.

Connect Grafana to InfluxDB & Share Team Dashboards

Once cAdvisor is writing to InfluxDB, you'll need to configure Grafana to connect to InfluxDB, then load up your team's shared Dashboards.

create-dashboareds.sh ensures that an InfluxDB database and user exists, that Grafana has a data store that points to it, and all Grafana dashboards are created.

Read create-dashboards.README.md for details.

Try It Out!

TODO: Update screenshots

Try out the "All Containers (Stacked)" dashboard, which shows the CPU, memory, and network activity for all of your Docker containers.

Note about Container Network Usage: if you use --net=host option in Docker, each container is reporting all traffic on the shared network interface, so you'll lose per-container visibility. You'll need to click on a single container to see the proper sum of all.

  1. Start your monitoring Docker cluster. The first time you run this might take a little while, since Composer will need to download Docker images from the Docker Hub:

    docker-compose up

  2. Wait 10 seconds, so the services have time to start.

  3. Run this command, which will set up Grafana to talk with InfluxDB, and create dashboards for any existing Docker containers:

    ./create-dashboards.sh

  4. View your new Grafana dashboard at http://localhost:3000 with admin/admin

Grafana screenshot

  1. View your new cAdvisor realtime performance monitoring dashboard at

http://localhost:9090.

cAdvisor screenshot

  1. Query for specific metrics using your new InfluxDB instance at http://localhost:8083.

InfluxDB screenshot

  1. Re-run the shell script at any point, to create dashboards for all of the currently-running Docker containers

    ./create-dashboards.sh