I am giving you the scenario as follows: We are deploying the build in Cloud foundary container(IASS) along with New relic binding services. This is hosted in Asia -South-east. My jmeter resides in same location but it is in Aazon ECM2-(asia south-east). While I ran the jmeter , seems response time looks higher compare to my New relic APP response time. Why soome time so much variation comes? is it due to latency factor? how to give explanation to my clientwhile they check New relic and Jmeter both the result. i am sure both are correct and need to find out rCA. Please helpenter image description here
JMeter response time includes network latency which we can not avoid. So, it might be because of latency. If it is a huge difference, try to run the test from a machine which is very close to the app server / same data center and see if it helps to minimize the latency.
What is the max no of users you are trying to simulate? What is the CPU, memory utilization of the load generator - usually i would keep them below 80%.
Ensure that your results are satisfying the little's law! check below for more details. if it does not satisfy you are trying to simulate too much load from your load generator which it could not handle - go for distributed load testing in that case.
http://www.testautomationguru.com/jmeter-distributed-load-testing-using-docker-rancheros-in-cloud/