Reportit: function for faling below threshold

[torstentfk]

Hello,

any solution for this problem:
I have a DS which is normaly 1 and in an error case it takes the value -2.
I want to have the time/duration of all values which are below 0.

Is reportit the right tool/plugin?

Torsten

[browniebraun]

Due to the fact that the sum of all values over zero is equal to the number of measured values which are 1,
you should be able use following formula to get the number of seconds:

f_num*step*(100-f_dot(0))/100


Best regards
-Andreas-


BTW: Have you made some pictures at the last CCC?

[NetAdmin]

I'm trying to do something very similar where I have a data source that checks the status of an interface and returns a value of 1 if the interface is up and 2 if the interface is down. So, I'm basically trying to build a report that will show how long (in minutes since that is the polling cycle) interfaces were down (or had a value of 2) during a given time period. Can't quite figure out the formula.

Regards,

John

[browniebraun]

I'm trying to do something very similar where I have a data source that checks the status of an interface and returns a value of 1 if the interface is up and 2 if the interface is down. So, I'm basically trying to build a report that will show how long (in minutes since that is the polling cycle) interfaces were down (or had a value of 2) during a given time period. Can't quite figure out the formula.

Regards,

John

Hey John,
in that case I would use "f_num*step*f_dot(1)/100/60".

Explanation: f_num*step will be the reporting period in seconds. f_dot(1) returns the percentage of all measured values greater than 1, so that "f_num*step*f_dot(1)/100" will return the number of seconds the interface was down.

Best regards
-Andi

[NetAdmin]

This is excellent, Andi.. Thanks so much.

Would it be possible to count the number of times that the value changed from 1 to 2? In other words, in addition to seeing how many total minutes the interface was down, we would like to see how many times it went down so that we can determine if it was a single outage or multiple outages. I hope this makes sense.

John

[browniebraun]

Not yet.
Sounds to me that we need one or two new functions.

Regards
-Andi

[torstentfk]

Hi,

first: I have no pictures from 5.eu.CCC - I am on the photos.

second: I tried this formular and it works. But be aware that you might have to change the rra to keep the granularity on the values.

For the additional formular, nice to have:
F_dbt (below threshold); or change it in:
f_tt (thold, direction): triggering threshold, direct=+1 -> over threshold; -1 = below threshold.
f_mx (thold, direction): maximum intervall of broken thold in seconds.


Torsten

[NetAdmin]

Not yet.
Sounds to me that we need one or two new functions.

Regards
-Andi

That would be great, Andi! I'd be more than happy to test out new functions prior to the next release..

John

[alkman82]

Let's get back to life this thread!
My purpose is to locate bottlenecks in our network. For example one Gigabit interface that has a flatted (haircuted) traffic for some time at 300Mbps. Iike the following image:

For the moment we have created a measurant with the following function "f_num*step*f_dot(f_max-100000)/100/60 "but seems wrong ... Any ideas please?