It’s all about the… containers!
summary
Explains how to monitor system containers (LXC) and application containers (Docker/Rancher/Kubernetes) using Icinga 2 with check_lxc and check_rancher2. Shows LXC limits: host /proc hides real memory/CPU until lxcfs is mounted, so CPU needs cpuacct diff and memory must come from cgroups; offers check_lxc that reads cgroup memory/swap and warns about inaccurate "available" values. Describes a Rancher production run with >1000 containers and a frozen pod stuck in removing state discovered via the Rancher API and fixed with kubectl.
AI-written summary · about
chapters (12)
- 0:00Intro
- 2:25That's a container
- 3:58The world is not perfect
- 7:44LXC - Linux Containers
- 9:45LXC - Monitoring processes
- 14:17LXC - Monitoring CPU usage
- 18:49Application Containers (AC)
- 20:51AC - Monitoring challenges
- 22:20AC - Management w/ Rancher
- 28:25AC - Using check rancher2
- 31:06Application Containers - Recap
- 32:41References and links
related talks
45mTroubleshooting containerized applications
01h 01mOrchestrating Docker Containers with Rancher
42mAutomating Infrastructure Deployment for Kubernetes
35mWho's Running My Pods? A Deep Dive into the K8s Container Runtime Interface
41mLeverage LXD/LXC with Kubernetes
30mAfterlife tales: Troubleshooting containerized applications
01h 39mLab: Linux Container Internals
13mPerformance and Energy Efficiency in Docker
