Container Basics On Linux (Namespaces/Cgroups)

Stop letting your container processes run wild on the CPU

🧩 The Challenge

You know that feeling when one misbehaving container starts eating all your CPU cycles and drags the whole node down with it? It’s infuriating when you’ve got ten other critical services on the same box just sitting there waiting for crumbs.

💡 The Fix

Use cgroups v2 to enforce a hard ceiling on how much processor time a group of processes can grab. It’s the easiest way to prevent a noisy neighbor from ruining your day.

echo 50000 > /sys/fs/cgroup/my_app/cpu.max

⚙️ Why It Works

Writing to the cpu.max file sets a quota of 50ms every 100ms period, effectively pinning that process to half a core regardless of how much work it tries to spin up. The kernel handles the throttling automatically once you drop that limit in.

🚀 Pro-Tip: Always check your current cpu.max before tweaking so you don’t accidentally starve a process that actually needs the burst capacity.

Linux Tips & Tricks | © ngelinux.com | 9/22/2026