Performance Graph reports inaccurate fractions

[rodpayne]

Problem: Performance Graph often reports count of, for example, "5.9067" instead of "6.0000".

Service Overview shows:

Code: Select all

Terminal Services Active = 6.00 Sessions

Advanced Status Details shows:

Code: Select all

Performance Data: 'Terminal Services Active'=6.000000Sessions;70.000000;100.000000; 

But, legend on Performance Graph shows:

Code: Select all

Terminal_Services_Active   5.9067 Sessions LAST    5.9067 Sessions MAX  ...

It seems to be doing some kind of internal floating-point rounding that gives an inaccurate number when converted back to decimal. If these numbers were rounded to the nearest integer, it would show the right numbers.

Version: Nagios XI 2012R2.2 downloaded with 64-bit Microsoft virtual machine.

Service settings:

Check command

Code: Select all

check_xi_service_nsclient
$USER1$/check_nt -H $HOSTADDRESS$ -s "$ARG1$" -p 12489 -v $ARG2$ $ARG3$ $ARG4$

$ARG2$

Code: Select all

COUNTER

$ARG3$

Code: Select all

-l "\\Terminal Services()\\Active Sessions","Terminal Services Active","Sessions" -w 70 -c 100

[abrist]

RRDs are database files that are a lossy aggregate. Values are averaged together over time. This is WAI.

[rodpayne]

Okay, I can see that for the AVERAGE value, but why would it be averaging or aggregating for the LAST and MAX values? Shouldn't they be the last value that was reported and the maximum value that has been reported? Is there any way to tell it to report in whole numbers?

[abrist]

Is there any way to tell it to report in whole numbers?

Unfortunately, not with RRDs. You may be able to modify the template to delay consolidation, but this will adversely affect all the other rrds generated from the template and lead to a large increase in rrd size. You will additionally need to recreate all your rrds, essentially losing the current historical data.