PRTG Manual: Calculating Percentiles
PRTG not only monitors your network and informs you in the case of issues that are worth a closer look, it also stores a lot of historic data that it gathers from your sensors. This means that you have a base for the statistical analysis and evaluation of what is and was happening in your network. When you create a report or a historic data report, you get raw data, sums, averages, and percentages of your monitoring data.
Additionally, PRTG also offers percentile calculation. This statistical method arranges your data, for example, from the lowest value to the highest value, and calculates the percentile that you want, optimally informing you about the distribution of your network-relevant data.
For example, if you request the 95th percentile, you know that 95 percent of the measured data is below a certain value and PRTG can tell you what this certain value is.
If applied to bandwidth, for example, you know which values you have when talking about the 5 percent of unusually high bandwidth consumption, and which value your users do not exceed 95 percent of the time. Service providers often use percentiles to offer billing that excludes infrequent usage peaks.
If you want to know more about the formula that PRTG uses for percentile calculation, see the Knowledge Base: What are percentiles and what differences do they make in PRTG reports?
What are percentiles and what differences do they make in PRTG reports?