'Maximum' value is often incorrect. Why?

mactfines · Post by **mactfines** » Mon Jul 24, 2006 7:34 am

Hi,

I am not sure whether the problem is my misunderstanding of rrdtool or Cacti, but it appears that Maximum values do not get displayed correctly.

Look at the first graph below, which states the Maximum is 28.97. Now the 2nd graph, where I clicked on the magnifying glass tool and zoomed in to just that day of the high spike, Saturday. It still states the Maximum is 28.97.

But look at the last graph, where I again zoomed in, from the 2nd graph, to just a 3-4 hour window. There I see the maximum is 55.99 (56).

The thing is, I know that 56 is the actual value. I watched this spike happen on Saturday, and still have the log data.

Why is the true Maximum not reflected on the graph until I zoom in, TWICE?

Thanks,
Ted

mdurand · Post by **mdurand** » Mon Jul 24, 2006 8:47 am

hi

i looked at my graphs to see if i've got the same problem
and i have

but i see that i only have my 'true maximum' for 5 minutes
and if i go back to the graph (without zoom) the maximum value is 'the false'

i'm not professional but i think when there's a too little spike cacti doesn't take care of it

you have to see it and zoom if you want to know

neurotox · Post by **neurotox** » Mon Jul 24, 2006 5:51 pm

i'm almost sure is because of the "space" between each data collection...

When you are max zoom... you see every 5 minutes data but if you check 1 week for example, you will see every hours data (that is only an example)

Post by **gandalf** » Tue Jul 25, 2006 1:01 pm

This is how consolidation works. Maxima will change in size, when the time they are covering increases. See more details in the first few links of my signature
Reinhard

Post by **TheWitness** » Tue Jul 25, 2006 8:54 pm

However, if you add the MAXIMUM, or it's installed for the Data Source, what I like to do is add a line that polls the MAXIMUM value instead of CUR or AVERAGE. Then you see the average for the range and the max for that same range.

TheWitness

Post by **gandalf** » Fri Jul 28, 2006 10:30 am

I do agree perfectly (suppose I stated sth like this in my docs?) You will notice the effect mostly on the weekly, monthly and yearly stuff (the daily does not show consolidation effects with standard rra setup).
See attached graph as an example
Reinhard

Post by **TheWitness** » Fri Jul 28, 2006 1:11 pm

I have to read those more often. Very informational

Larry

Post by **gandalf** » Sat Jul 29, 2006 5:05 am

I'm discussing this stuff with peoples around for _ages_

Reinhard

'Maximum' value is often incorrect. Why?

'Maximum' value is often incorrect. Why?

Who is online