Shulou(Shulou.com)06/02 Report--

This article focuses on "the implementation of docker container monitoring". Interested friends may wish to have a look at it. The method introduced in this paper is simple, fast and practical. Let's let the editor take you to learn "the implementation of docker container monitoring".

This article was written in 2015, and all the content related to the PAAS platform has been completed in 2015Q3. At that time, the version of docker used was 1.6.2. Although the new version of docker was released quickly, the monitoring content mentioned below is roughly the same.

1. What metrics do you need to monitor in docker containers?

Container CPU, memory, IO, network, application survival

Second, how to monitor?

Host monitoring zabbix can be realized, and zabbix is installed inside the container.

(the data is consistent with the host and cannot reflect the real status of the container)

See how others achieve it?

Searching GitHub...

Seagull

You can view and start / stop containers and manage images, but you cannot view the container's use of CPU, memory, IO, network, and so on.

Prometheus

Too big

Docker official stats API

Docker version 1.5 or above is required. There are two ways: command line and API. API is non-interruptive and constantly displays data.

Docker-mon

Command line tool, version 1.5 or above, the display effect is very good, but the data cannot be summarized and the historical data cannot be viewed.

Dockernana

You can monitor the daily cpu, memory and IO of the docker host, but you cannot monitor the resource usage of the container.

CAdvisor

You can check the container's use of CPU, memory and disk, as well as the WEB interface. The disadvantage is that each host needs a container, and the data cannot be summarized and the historical data cannot be viewed.

Figure 1:Docker-mon

Figure 2:cAdvisor

The final script is as follows:

A1roomuser= `cat / cgroup/cpuacct/docker/$i/cpuacct.stat | grep user | awk'{print $2}'`

A1roomsys = `cat / cgroup/cpuacct/docker/$i/cpuacct.stat | grep sys | awk'{print $2}'`

Sleep 5

A2roomuser= `cat / cgroup/cpuacct/docker/$i/cpuacct.stat | grep user | awk'{print $2}'`

A2roomsys = `cat / cgroup/cpuacct/docker/$i/cpuacct.stat | grep sys | awk'{print $2}'`

Cpuall=$ ((a2_user + a2_sys-a1_user-a1_sys))

CPUSTAT= `echo "scale=2;$cpuall/5/100" >

Memory

# @ currently used memory divided by maximum memory limit

[root@BJ-YF-64-7] # cat / cgroup/memory/docker/$DOCKERID/memory.limit_in_bytes

1073741824

[root@BJ-YF-64-7] # cat / cgroup/memory/docker/$DOCKERID/memory.usage_in_bytes

186269696

Network traffic

# @ obtain the network status of the process according to the container PID

PID= `docker inspect-f'{{.State.Pid}}'$i`

A1roomrx = `cat / proc/$PID/net/dev | grep eth2 | awk'{print $2}'`

Number of network connections

# @ obtain the network status of the process according to the container PID

Cat / proc/$PID/net/tcp

4. Preliminary display of shell+InfluxDB+Grafana

Print the results using Shell:

Use InfluxDB+ Grafana to show the results:

5. Connect to Zabbix

If the problem of data display is solved, what about the data alarm? Zabbix

Zabbix a host has N pictures, container 1CPU, container 1 memory, container 2CPU.

Data acquisition is complex and difficult.

Dismantle

Each container runs Zabbix

The data is printed to the container's mapping directory in the host through the shell script.

Historical trend chart

At this point, I believe that everyone on the "docker container monitoring implementation method" have a deeper understanding, might as well to the actual operation of it! Here is the website, more related content can enter the relevant channels to inquire, follow us, continue to learn!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.