Skip to main content

Service to perform automatic shutdown of idle container hosts

Project description

Docker Idle Watchdog

The docker-watchdog Python package provides a background service that monitors Docker container hosts for periods of inactivity and performs an automatic shutdown of the host after a predefined idle threshold is reached. In a similar manner to the shutdown-if-idle package, the intended use case is the automatic shutdown of cloud VMs that have been started on demand as part of a Continuous Integration (CI) pipeline, albeit with a focus on Docker container hosts.

The key features of the package include:

  • Supports Windows, macOS, and Linux
  • Provides easy service installation under all supported operating systems
  • Takes cloud provider billing granularity into account to ensure shutdown decisions are cost-effective




The docker-watchdog package requires the following:

  • Python 3.5 or newer
  • Under macOS and Linux, the pleaserun Ruby gem needs to be installed to perform service installation

Package installation

To install the docker-watchdog Python package itself, simply run:

pip3 install docker-watchdog

(You may need to prefix this command with sudo under macOS and Linux.)

Once the package is installed, you can run the watchdog via either of these commands:

# Uses the wrapper generated by pip

# Invokes the package directly via the Python interpreter
python3 -m docker_watchdog

Service installation

To install the startup service for docker-watchdog, ensure you have any necessary dependencies installed and run the following command with elevated privileges:

docker-watchdog --install

The background service will then start automatically the next time the host system boots up.

Installation scripts for cloud servers

To simplify the setup process for cloud servers, the scripts directory contains installation scripts for common Docker container host platforms. The scripts automatically install docker-watchdog and its dependencies, and register the docker-watchdog startup service.

Scripts are provided for the following platforms:

  • Windows Server: scripts/install_windows_server.ps1

    Installs the Chocolatey package manager, uses Chocolatey to install Python and docker-watchdog, and registers the docker-watchdog startup service.

    To use this script, run the following command from an elevated PowerShell prompt:

    Set-ExecutionPolicy Bypass -Scope Process -Force; iex ((New-Object System.Net.WebClient).DownloadString(''))
  • Debian-based Linux distros: scripts/

    Installs Python, Ruby, pleaserun, and docker-watchdog, and registers the docker-watchdog startup service.

    To use this script, run the following command with root priveleges:

    curl -fsSL '' | bash


The watchdog can be configured either via a JSON configuration file or via environment variables. These two methods can also be used together, in which case values from the configuration file will take precedence over values from environment variables. Any settings for which a value has not been specified will fallback to using a sane default.

The location of the JSON configuration file is based on the operating system and runtime environment:

  • Under Windows: %APPDATA%\docker-watchdog\config.json
  • Under macOS and Linux:
    • If running as a non-root user: $HOME/.config/docker-watchdog/config.json
    • If running as the root user: /etc/docker-watchdog/config.json

The available configuration settings are as follows:

JSON Key Environment Variable Default Description
sleep DOCKER_WATCHDOG_SLEEP_INTERVAL 60 Specifies the interval (in seconds) to sleep for between sampling runs.
timeout DOCKER_WATCHDOG_IDLE_TIMEOUT 600 Specifies the period of inactivity (in seconds) required to consider the system idle.
billing DOCKER_WATCHDOG_BILLING_GRANULARITY 0 Specifies the billing granularity (in seconds) if the host is a cloud VM, or zero otherwise.
percentage DOCKER_WATCHDOG_EFFECTIVE_PERCENTAGE 0.9 Specifies the minimum percentage (0.0 to 1.0) of the current billing unit that must have elapsed to consider a shutdown to be cost-effective.


Copyright © 2019, Adam Rehn. Licensed under the MIT License, see the file LICENSE for details.

Development of this package was funded by Deepdrive, Inc.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release. See tutorial on generating distribution archives.

Built Distribution

docker_watchdog-0.0.5-py3-none-any.whl (10.8 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor NVIDIA NVIDIA PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page