Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials

95th Percentile Discrepency - Text v Chart

jbaulsir

On this chart, why does the text say 95th Percentile 71 Mb/s and the line on the chart is almost at 75 Mb/s?

ChartCA1RRNPA.jpg

Find more posts tagged with

Accepted answers

All comments

mark_wiggans

Here is an explanation http://www.solarwinds.com/NetPerfMon/SolarWinds/OrionPH95th.htm

jbaulsir

Thanks Mark, but I understand how it's calculated, my question is why there is a difference between the caculated number and the position of the line on the chart.

FormerMember

I'll look into that. I can tell you that we're replacing all the chart controls in the next version, so any discrepancy will be addressed there.

Network_Guru

I brought up a similar issue several years ago related to 95th percentile calculation.
I'm not sure this has ever been addressed.

What happens when the calculation is done over an extended period?
For example, if your nightly stats rollover settings are set to:
>detailed to hourly after 7 days
>hourly to daily after 14 days

Your last 7 day chart will be accurate if set to an interval equal or greater than your polling interval, but a last 30 day chart will not.
Assuming you set the last 30 day chart to an hourly interval then the first 14 days will have 336 samples but the last 16 days will only have 16 samples for a total of 352 samples. Now throw out the top 5% (18 samples) to get 95th percentile, and you can see the problem with this logic.

My issue at the time was with %availability stats using report writer.
Don Yonce had mentioned that Report Writer would not calculate this correctly, but the logic built into the charts and graphs was correct. Can someone from SW confirm this?