Check the disk io between my VM and another VM using telegraf(clearcase as an example)

1

Hey,

So how would I get the metrics of disk io (read and write) between 2 servers? I know how to do it for my current server and its local disk, but confused on how to set it up to another server?

Any help would be appreciated!

2

Hi @Tyler_D

If you just want to collect metrics from the server, then install the Telegraf agent on the servers you want to monitor and make sure the telegraf config file points to your influx instance and can authenticate with influxDB. After a couple of minutes you should see data from the other servers arrive in the database.

3

how do i solve this issue?

cant seem to get the disk io to work
E! [inputs.diskio] Error in plugin: error getting disk io info: context deadline exceeded

4

Is this a windows or Linux server you want to monitor?

5

windows unfortunately

6

Could you try with the windows performance counters plugin?

i think you want the win_physicaldisk measurement for disk io

7

win_diskio ?

What is the difference between that and diskio?

8

The main difference is the inputs.diskio, inputs.cpu and so on use WMI - The win_perf* counters are more efficient and the recommended option for monitoring Windows servers. I only know that because its in the config, i couldn’t find any documentation referencing it though.

Other than that, I’m unsure what other differences there are. Generally though I would use the inputs.cpu type inputs for monitoring Linux machines. Win_perf for windows servers.

9

ok, thank you for your help!

side note, what is WMI>
?

10

Windows Management Instrumentation.

11

what measurements does the win_diskio use? are they measured in ms, iops, etc?
Just want to know what the best measurement is to monitor it in Grafana?

currently using this, but not sure if thats the best way to monitor it?

SELECT mean(“Percent_Disk_Read_Time”) FROM “win_diskio” WHERE (“host” = ‘xxxxx’ AND “objectname” = ‘PhysicalDisk’ AND “instance” = ‘0 C:’) AND timeFilter GROUP BY time(__interval) fill(null)

like the percent disk read time…normal disk io its just read time and the units are in milliseconds (ms) … just not sure what the percent disk portion indicates, if its either the same, but in percent form?