The Linux kernel’s container control group mechanism is called cgroups, short for control groups.
Cgroups let the kernel limit, account for, prioritize, and isolate resource usage by groups of processes. Controllers can manage resources such as CPU time, memory, disk I/O, CPU sets, and the number of open files. This makes it possible to keep one workload from consuming all of a host’s resources.
Google engineers began developing the feature in 2006 under the name “process containers.” The terminology changed to “control groups” in 2007, and the functionality entered the mainline Linux kernel around the 2.6.24 era. Docker, Kubernetes, LXC, and systemd all use cgroups, but they are users of the kernel mechanism rather than alternatives to it.
Cgroups are often discussed alongside namespaces. Namespaces isolate what processes can see—such as processes, mounts, or network interfaces—whereas cgroups control and measure resource consumption. Cgroup version 2, with a unified hierarchy, was merged in Linux 4.5 in 2016.