Missing Capacity Planning Graphs on Drives

[michal.nastaly]

Hi,

I have environments which are based entirely on passive checks. One of these check's is to monitor drives on the servers using check_drivesize command without specifying any drive. This means that the service check returns values for all the drives on the server. Because of that I don't see capacity planning graphs being drawn. If i change the command and specify drive i.e drive=c the capacity planning graphs appear.

The only message I can see on capacity planning screen is : "There are no capacity planning graphs available for this host."

The client plugin that we use for monitoring is NSClient++ 0.5.0.62 (newest)

The Nagios XI version at the site I am testing this config on is 5.4.2.

Regards
Michal Nastaly

[avandemore]

My best guess is that your plugin isn't returning any performance data.

What does the command and output look like when run from the CLI?

[tmcdonald]

If i change the command and specify drive i.e drive=c the capacity planning graphs appear.

This is suspect. Did you originally specify a drive when you set up the service, or did you do that later? Changing a check can have some interesting effects, especially in cases like this where you are adding or removing information the check is returning (in this case, number of drives). If the check originally expected information for just one drive, then the RRD file it made to contain that info only has room for that one check. If you later change this to include more drives, it will choke when trying to insert the data.

[michal.nastaly]

What does the command and output look like when run from the CLI?

Code: Select all

check_drivesize "exclude=F:" "warning=free<200g" "critical=free<50g"
L        cli CRITICAL: CRITICAL C:\: 253.418GB/447.031GB used, D:\: 0B/0B used
c:\Program Files\NSClient++/nsclient.log could not be opened, Discarding: message: CRITICAL: CRITICAL C:\: 253.418GB/447.031GB used, D:\: 0B/0B used
L        cli  Performance data: 'E:\ free'=1.76719TB;0.19531;0.04882;0;1.81935 'E:\ free %'=97%;11;3;0;100 'C:\ free'=193.61344GB;200;50;0;447.03124 'C:\ free %'=43%;45;11;0;100 'D
:\ free'=0GB;200;50;0;0
c:\Program Files\NSClient++/nsclient.log could not be opened, Discarding: message:  Performance data: 'E:\ free'=1.76719TB;0.19531;0.04882;0;1.81935 'E:\ free %'=97%;11;3;0;100 'C:
\ free'=193.61344GB;200;50;0;447.03124 'C:\ free %'=43%;45;11;0;100 'D:\ free'=0GB;200;50;0;0

[michal.nastaly]

If i change the command and specify drive i.e drive=c the capacity planning graphs appear.

If you later change this to include more drives, it will choke when trying to insert the data.

Is there a way to fix this?

[michal.nastaly]

If i change the command and specify drive i.e drive=c the capacity planning graphs appear.

Did you originally specify a drive when you set up the service, or did you do that later?

It was new passive service check, never specified drive.

[michal.nastaly]

Code: Select all

[/settings/scheduler/schedules/Drives]
command = check_drivesize "exclude=F:" "warning=free<200g" "critical=free<50g"
interval = 1m


[/settings/scheduler/schedules/Drive]
command = check_drivesize "drive=C:" "warning=free<200g" "critical=free<50g"
interval = 1m

the top check - capacity planning doesn't work
the bottom check - capacity planning works

i have drive c:\ and e:\

[michal.nastaly]

The error message that I am getting on the capacity planning grap is:

Code: Select all

Attribute Error - 'NoneType' object has no attribute 'something'

I'm starting to believe that there is nothing wrong with the configuration and its the capacity planning tool that is unable to provide graphs for monitoring multiple drives with a single command.

[tmcdonald]

After speaking with the developers this sounds like the issue is caused by switching from single-drive checks to multiple, and vice-versa as I suspected. Unfortunately the way around this is to specify the drives individually, or else ensure that the number of drives does not change if you are checking multiple in one check.

I've asked about the feasibility of storing each separate perfdata result in its own RRD, so that may be an option in the future.

[michal.nastaly]

Hi Guys,

This is suspect. Did you originally specify a drive when you set up the service, or did you do that later? Changing a check can have some interesting effects, especially in cases like this where you are adding or removing information the check is returning (in this case, number of drives). If the check originally expected information for just one drive, then the RRD file it made to contain that info only has room for that one check. If you later change this to include more drives, it will choke when trying to insert the data.

After further investigation I can confirm that @tmcdonald is right. I must have change the command since it was added to the config and nagios xi. After removing the RRD file on one of the hosts and re adding the check the graphs appeared.

Here comes another question: I used to have passive check called "Drive C" which used to monitor the C drive. Then I have added new one called "Drives" and removed the "Drive C". Because of this we have lost all the graphs that we had (12 months +). The Drive C rrd files are still on the nagios xi server in the host folders. When I add the Drive C check back in to the config how can I make sure it will pickup those files and restore the graphs that we had for last 12 months?