How application traffic drops is calculated?

[Yosi_Neuman]

Hi guys,

We have set the detect traffic drops to 100% over 5 minutes

but problem was fire for a drop from 49.9 to 19.4

We followed the formula 

And its look like a decrease of 61% 

Which is less than 100% .... wonder why problem was fired?

Is there any other explanation on how traffic drops are calculated?

Thanks in advance 

Yos  

[Peter_Youssef]

Hello @Yosi_Neuman 

Regarding the raised inquiry:

1. Dynatrace application traffic anomaly detection is based on the assumption that most business traffic follows predictable daily and weekly traffic patterns.
2. Dynatrace automatically learns each application's unique traffic patterns. Alerting on traffic spikes and drops begins after a learning period of one week because baselining requires a full week’s worth of traffic to learn daily and weekly patterns.
3. Following the learning period, Dynatrace forecasts the next week’s traffic and then compares the actual incoming application traffic with the prediction. If Dynatrace detects a statistically significant deviation from forecasted traffic levels, it raises an alert.

Related Ref:

• https://docs.dynatrace.com/docs/shortlink/automated-baselining 

KR,

Peter.

[Peter_Youssef]

Hi @Yosi_Neuman 😊

If we are seeking the exact formula 😎 developed by Dynatrace 😉, the best source of information would be Dynatrace product expert through in-Product expert live chat or Raising support ticket.

I had Infra related calculation any I had gotten the insightful information.

All the best.

BR, 

Peter

[rastislav_danis]

Hi @Yosi_Neuman, where did you get that formula? I'm not able to find it in docs. I can see 9 unexpected low traffic events in your problem. All of them are with same/similar percentage as in picture?

[Yosi_Neuman]

Hi @rastislav_danis 

Its not from dynatrace, its just a formula I have found with google help. 😆

Still wonder what is dynatrace formula 🤔

Yos  

[Yosi_Neuman · Accepted solution]

OK, we got answer from support / R&D!

The formula is: Alert if Observed Traffic is Less than Expected Traffic Percentage

Here is an example of how it works

HTH

Yos 

[rastislav_danis]

So in your case for each 1 minute interval of 5, actual value was lower than expected value. Wondering if mentioned 19.4 vs 49.95 is average of those 5 intervals or just one of them (first, last, ..).

[Yosi_Neuman]

Hi @rastislav_danis 

From the label above minutes field its look like the observed traffic has to be below for the whole period of time.  

HTH

Yos

[rastislav_danis]

That's clear, but each minute can have different "expected traffic" value. So what i ask (myself 😉 is: drop from 49.95 to 19.4 happened in first minute of 5 minute interval?