From 4a9418db8c622eb291e7b9860f440ceac13e463a Mon Sep 17 00:00:00 2001 From: Nihal LK Date: Wed, 9 Oct 2024 14:35:19 +0530 Subject: [PATCH] Updated ReadMe file as per review Signed-off-by: Nihal LK --- performance-test/README.md | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) diff --git a/performance-test/README.md b/performance-test/README.md index b3d63830a7..b51fbbc734 100644 --- a/performance-test/README.md +++ b/performance-test/README.md @@ -101,20 +101,24 @@ https://mosip.atlassian.net/wiki/spaces/PT/pages/1227751491/Steps+to+set+up+the+local+system#PluginManager ### Designing the workload model for performance test execution + * Calculation of number of users depending on Transactions per second (TPS) provided by client +* The script and the below calculation is preconfigured as per 100 tps, if you are testing for other tps, the below values neds to be adjusted. + * Applying little's law * Users = TPS * (SLA of transaction + think time + pacing) * TPS --> Transaction per second. * For the realistic approach we can keep (Think time + Pacing) = 1 second for API testing * Calculating number of users for 10 TPS - * Users= 10 X (SLA of transaction + 1) - = 10 X (1 + 1) - = 20 + * Users= 100 X (SLA of transaction + 1) + = 100 X (1 + 1) + = 200 ### Usage of Constant Throughput timer to control Hits/sec from JMeter -* In order to control hits/ minute in JMeter, it is better to use Timer called Constant Throughput Timer. + +* In order to control hits/ minute in JMeter, it is better to use Timer called Constant Throughput Timer. This is calculated explicitly for each thread group based on the scenario's weightage * If we are performing load test with 10TPS as hits / sec in one thread group. Then we need to provide value hits / minute as in Constant Throughput Timer * Value = 10 X 60