OOMKilled

An OOMKilled event occurs in computer systems, particularly within operating systems like Linux, when a process is terminated due to an out-of-memory (OOM) condition. The term originates from the Linux kernel's Out-of-Memory (OOM) killer, a subsystem designed to manage memory shortages by forcibly terminating low-priority processes when physical RAM is exhausted. This mechanism prevents the system from crashing by ensuring critical services remain operational.

The OOM killer prioritizes processes based on factors such as memory usage, fairness (e.g., fairness to system daemons), and user space considerations. It typically targets processes with the highest memory consumption, though it may also consider other criteria like process age, nice values (priority adjustments), or cgroup (control group) constraints. When invoked, the OOM killer logs the event in the kernel ring buffer, often visible via tools like `dmesg` or `journalctl`, where it may appear as a message indicating the process's PID (process identifier), name, and reason for termination.

OOMKilled events are generally undesirable as they disrupt running applications and may lead to data loss or degraded performance. Common causes include excessive memory allocation by a process, memory leaks, or insufficient system resources to accommodate all active workloads. Developers and system administrators often investigate such events to identify memory-intensive processes, optimize resource usage, or adjust system configurations to mitigate future occurrences. Tools like `top`, `htop`, or `valgrind` can help diagnose memory-related issues contributing to OOM conditions.