1. Practice of Monitoring Docker Container
In Data Collection, we introduced the common data sources. As a monitor framework, Open-Falcon can collect the monitor index data of any system as long as they are converted to the standard format of Open-Falcon.
The data collection of docker container can be done through micadvisor_open.
1.1. Working Principle
Micadvisor-open is a module for monitoring the resource of docker container based on Open-Falcon. The data of CPU, memory, diskio and networkio will be pushed to Open-Falcon after being collected.
1.2. Collected Index
Counters | Notes |
---|---|
cpu.busy | percentage of cpu usage |
cpu.user | user-mode percentage of cpu usage |
cpu.system | kernel-mode percentage of cpu usage |
cpu.core.busy | usage of each cpu |
mem.memused.percent | percentage of memory usage |
mem.memused | amount of memory usage |
mem.memtotal | total amount of memory |
mem.memused.hot | hot usage of cpu |
disk.io.read_bytes | bytes diskio reads |
disk.io.write_bytes | bytes diskio write |
net.if.in.bytes | incoming bytes of networkio |
net.if.in.packets | incoming packets of networkio |
net.if.in.errors | incoming errors of networkio |
net.if.in.dropped | incoming droppings of networkio |
net.if.out.bytes | outgoing bytes of networkio |
net.if.out.packets | outgoing packets of networkio |
net.if.out.errors | outgoing errors of networkio |
net.if.out.dropped | outgoing droppings of networkio |
1.3. Contributors
- mengzhuo: QQ:296142139; MAIL:mengzhuo@xiaomi.com
1.4. Complementary Information
- The lib database collected by another docker metric: https://github.com/projecteru/eru-metric