thold baseline deviation value

[_raindrop18]

Greeting!!
I want to add my advance ping to thold 4.1. my objective is to create a baseline monitoring. what I am looking for the get the alert when the latency double for current baseline value.

this is my current value but appears to be the result is not correct. it's not alert when it's double but on much lower value.

Reference in the past - 604800 (week)
Time range - 86400
Baseline deviation UP - 200 %
Baseline deviation UP - 0
Baseline Trigger Count - 2

did i miss anything? any suggestion. thanks

[noname]

On thold-0.4.3 or prior, current value is compared against (maximum + up%) and (minimum-down%) during the given time range.
- http://forums.cacti.net/viewtopic.php?p=206719#p206719

[_raindrop18]

Thanks much! I am glad I got a response from you, because I have seen your nice drawing about thold but I didn't figure out what exactly mean. would you mind to confirm if i understood correctly this time around
my objective is to get alert when the ping latency double the baseline ms

Code: Select all

Reference in the past - 604800 (week)
Time range - 86400
Baseline deviation UP - 50%

example :- baseline (50) + up (50% ) = 100

so can get the alert when the base line calculation reach 100 ms? is the correct ?????

thanks again for your help!!!!

[noname]

As you know, baseline triggering is used to detect "abnormal" changes.
"Baseline" (nearly equal to "normal") value is determined by calculating from historical data.

Thold-0.4.3 or prior: "baseline" = Values ranging minimum to maximum in the given time range
Thold-0.4.4(maybe) or later: "baseline" = An average value in the given time range

example :- baseline (50) + up (5000% ) = 100

so can get the alert when the base line calculation reach 100 ms? is the correct ?????

"50" is the maximum latency which you can accept?
Or the latency which have been usually observed? (If so, please upgrade thold plugin)

Anyway, if you want to get alert when current value was doubled from baseline value, try to set "100" (%) to Up Deviation.

e.g.) 50ms + 50ms * 100% = 100ms

[_raindrop18]

"50" is the maximum latency which you can accept?
Or the latency which have been usually observed? (If so, please upgrade thold plugin)

my objective is when the latency double from normal range. we have remote location all over the country, so the latency from nearest location are normally around (50ms) but the farthest one is around (110ms). so I can't put one fixed number that's why I want to get alert when the normal latency reach double.
so I should upgrade thold to the newer one? since I have some modified script and not even sure the newer one is compatible with my cacti v 8.e and P.A 2.5. those are the reason I have kept thold 4.3 but if my objective can't be done without upgrade then I will upgrade it...what do you think?

thank you so very much for your time!!!

[noname]

>> so I should upgrade thold to the newer one?

Thold-0.4.3 or prior: "baseline" = Values ranging minimum to maximum in the given time range
Thold-0.4.4(maybe) or later: "baseline" = An average value in the given time range

If there is a little difference between maximum and average latency, perhaps both versions will act similarly.
In other words, if you are not tolerant of fluctuation, you should use newer one.

[_raindrop18]

ok, I will try the newer version. thanks a lot!!!