How to configure CPU and RAM resource limits for Docker containers?

Limiting system resources for containers is a critical aspect of infrastructure management in Docker. By default, running containers can use an unlimited amount of CPU time and RAM on the host machine, which in the event of a failure or memory leak in one of the applications can lead to a crash of the entire system.

To avoid such problems, administrators and developers can set hard limits at the container startup stage using special command-line flags or through orchestrator configuration files. This allows for efficient distribution of computing power among various services on a single server.

To manage RAM, the memory and memory-swap flags are used. The memory flag sets the maximum amount of RAM that the container can use, while memory-swap defines the total amount of memory and swap space. For example, a startup command with a memory parameter of 512 megabytes ensures that the application will not exhaust all host resources.

CPU management is implemented through shares and cpus mechanisms. The cpu-shares parameter sets the relative weight of the container compared to others when CPU resources are scarce, while the cpus parameter allows you to strictly limit the number of physical or virtual CPU cores available to a specific instance of the application.

To permanently apply such limits in a production environment, it is recommended to use a declarative approach with Docker Compose. In the configuration file for each service under the deploy and resources sections, you can specify exact limits and reservations for CPU and memory, thereby ensuring the stability and predictability of the entire microservice architecture.