Monitoring Ceph clusters with Telegraf+influxdb+grafana

Telegraf is a collection of monitoring agents, there are support to collect a variety of data plug-ins such as Ceph,apache,docker,haproxy,system, but also support a variety of output plug-ins such as Influxdb,graphite and so on.

Influxdb is a time series database and is used for monitoring scenarios

Grafana is a great drawing tool.

The combination of the three has three main processes:

& Nbsp;               1. The Telegraf installed on all nodes of the Ceph cluster (including the Mon and OSD nodes) collects Ceph cluster information through the Ceph_input plug-in, where two pits are stamped, and the Ceph_input plugin reports two errors, one error reading from Socket '/var/run/ceph/ceph-mon.control-1.asok ': Error running Ceph dump:exit status 22, and the other is error in input [Inputs.ceph]: Error executing command:error running ceph status:exit status 1. After the elimination of the first error is Telegraf bug, need to add the sock file W permissions to the Ceph process, the first error is because the user who executes the TELEGRAF process does not have ceph_client_admin read access to the keyring file. To solve these two problems, the Telegraf ceph_input plugin will work properly.

2.telegraf through the Ceph_input plug-in collection of ceph information through the Influxdb_ouput data written to the INFLUXDB,INFLUXDB database through the 8086 port to receive data, as if the UDP protocol is also supported.

3.grafana get data drawing from Influxdb.

Take a picture.

650) this.width=650; "Src=" Http://s3.51cto.com/wyfs02/M01/8A/E3/wKioL1g-mc3Rs0CNAAFFzg3QCz0844.png-wh_500x0-wm_3 -wmp_4-s_2114955627.png "title=" C.png "alt=" Wkiol1g-mc3rs0cnaaffzg3qcz0844.png-wh_50 "/>

Time to write a detailed point.

This article from "Tanyuhua" blog, reproduced please contact the author!

Monitoring Ceph clusters with Telegraf+influxdb+grafana