Hello,

it did first steps in adjusting a sensors warn and error level by analysing historic data. So I read about and started some historic data reports with different adjustments of percentile type, average interval and mode.

First, the outcome of the report was not clear to me. First it switched to calculating a medium value even though I said I want to see raw data. This happens even if I choose a very short time window just containing a few measures.

Second problem is, I could not see which values are used for calculating a medium value. So if my sensor has a interval of 60 seconds and I set the average interval to 180, which measures belong to the medium calculation.

Third, the impact of choosing medium, maximum or minimum mode of the channel to the outcome of the report was not explained.

I found nothing special in the documentation about that, just rough documentation.

Could anyone explain that.

Regards

Article Comments

Hello,
we appreciate your inquiry.

When querying for historic data within PRTG, the following averaging rule applies:

For performance reasons, PRTG automatically averages monitoring data when calculating data for large time spans. For periods of Up to 40 days you can use any average (even Raw(no average)), but for periods over 40 days PRTG will automatically set the average to 60 minutes/1 hour or larger.

Regarding how percentile works/behaves within PRTG, please check the following links:

Best Regards,


Jan, 2016 - Permalink