Heres another slide from the Baselining session I had the other day.
This time we were looking at application behavior. The point of this slide is not to get initially caught up with what could be causing the behavior, or if its bad or good.
In this specific example, I used wget to see if an HTTP transfer behaved the same way as when the application performed the same HTTP GET.
I kept every as consistent as possible:
• Same source IP
• Same destination IP
• Same version of HTTP (1.1)
• Same URL
• Same approx. time of the day (5 minutes apart)
During a real baseline I would want to get multiple traces/tests (5 if possible) to definitively prove if this is indeed a real pattern.
After proving the pattern is consistent, you can make an observation, possible recommendations and then remeasure to see if it helped.
...
https://www.youtube.com/watch?v=Qc9azidbesw