Discrepancy in PNP report

Support forum for Nagios Core, Nagios Plugins, NCPA, NRPE, NSCA, NDOUtils and more. Engage with the community of users including those using the open source solutions.
Locked
sattanathan
Posts: 23
Joined: Tue Oct 23, 2012 8:10 am

Discrepancy in PNP report

Post by sattanathan »

Hi,

I have attached cpu utilization PNP report for one of the server. I have seen in the yesterday report, the maximum usage of the cpu is 90.7956%. However, in the last week(including yesterday) report the cpu maximum value is showing only 90.3632%.

Please clarify why the value 90.7956% is showing in the weekly report as a maximum.

Regards,
Sattanathan.S
Attachments
pnp error.JPG
abrist
Red Shirt
Posts: 8334
Joined: Thu Nov 15, 2012 1:20 pm

Re: Discrepancy in PNP report

Post by abrist »

This is most likely due to how rrdtool stores and retrieves information from the rrds. As rrds are a binary aggregate, rrdtool averages values. Additionally, after a day, values will get averaged as they get moved farther back into the past in the rrd. You will find that the larger the time frame, the more averaging of the values are performed before the image is created. The more particular/granular the time frame, like the previous day, will have higher resolution in data points.
Former Nagios employee
"It is turtles. All. The. Way. Down. . . .and maybe an elephant or two."
VI VI VI - The editor of the Beast!
Come to the Dark Side.
sattanathan
Posts: 23
Joined: Tue Oct 23, 2012 8:10 am

Re: Discrepancy in PNP report

Post by sattanathan »

Thanks Abrist..

But if this is the case "data will be averaged for larger time frame" then how can i correlate the weekly/monthly/yearly and daily graphs and can attain similarity between the graphs.

Thanks,
Sattanathan.S
sreinhardt
-fno-stack-protector
Posts: 4366
Joined: Mon Nov 19, 2012 12:10 pm

Re: Discrepancy in PNP report

Post by sreinhardt »

What do you mean by similarity in graphs? When rrd compresses data, it takes the min, max, and average and averages each individual value over a period of time. So for instance if it was compressing 4 weeks of data into one month point, it would take the minimums for that period, average, then store as the month average, and so on for each of the other two values. The data is still very representative of the data over that time frame, it is just not as exact as one day measurements which might still be in individual checks or hourly averages.
Nagios-Plugins maintainer exclusively, unless you have other C language bugs with open-source nagios projects, then I am happy to help! Please pm or use other communication to alert me to issues as I no longer track the forum.
sattanathan
Posts: 23
Joined: Tue Oct 23, 2012 8:10 am

Re: Discrepancy in PNP report

Post by sattanathan »

Thanks Abrist & sreinhardt for clearing my doubt.

Regards,
Sattanathan.S
abrist
Red Shirt
Posts: 8334
Joined: Thu Nov 15, 2012 1:20 pm

Re: Discrepancy in PNP report

Post by abrist »

No problem. RRDs are a bit odd. But to store this data as flat metrics would consume disk space at an astronomical rate. It is one of the necessary evils of monitoring.
Former Nagios employee
"It is turtles. All. The. Way. Down. . . .and maybe an elephant or two."
VI VI VI - The editor of the Beast!
Come to the Dark Side.
Locked