A Docker container performance monitoring tool.

DockerMon is a tool to track the resource consumption of Docker containers over time. By default, the Docker API only exposes metrics instantly, which is only really useful during forensic examination or diagnosis of runtime issues. This tool is designed to be used during the development lifecycle of products, to track performance metrics and provide critical feedback about performance bottlenecks.

This project was started by Josh Griffith after finding that the majority of tools which do a similar job have severe limitations. Typically, they either exclude critical data from the endpoint or simply don't run cross platform. Docker is a cross platform tool, and provides a host of useful information. The intention of this tool is to make full use of those facts.

The tool is very simple. It polls docker stats regularly to extract metrics, then stores them in a database. A CherryPy server then serves these metrics via a RESTful API. Currently the API is limited in scope, but it intends to eventually serve data from time ranges (e.g start/end time), from time history (e.g seconds ago), or from any other reasonable time period. It's all pretty simple in that sense.

Just pip install -r the requirements, copy config.ini.sample and edit it to your own configuration, then run python ./main.py and hit the endpoint. By default it runs on port 8080. You can use the CLI to pass in parameters for the target host and database file instead of the config file.