Solved: Re: Performance testing for apigee with Loadrunner

sureshtirumalap · 04-04-2018 10:34 PM

Need Performance testing strategy for apigee using Loadrunner.

Can someone help me on providing some documentation?

Also I need compare the perf test results with JMeter Vs Loadrunner. Please help on this.

DChiesa

I think your strategy for testing APIs hosted in Apigee Edge is not different than the appropriate strategy for testing APIs hosted anywhere else. I don't think there is special documentation in this regard. Just invoke the APIs using the appropriate dynamic request content.

As for test execution:

use a suitable warmup period. Caches get filled, pages get filled.
vary connection times and payload sizes to model expected real-world transactions
try to model the number of API keys or tokens - you cannot run an effective test using only a single token if your real-world client base will use 5000 concurrent tokens.
run small tests for at least 10 minutes
conduct many different trials
run at least one long soak test - 48 hours or so

I also don't have any guidance for you in comparing tests implemented in JMeter versus tests implemented in LoadRunner. That seems like it ought to be simple to do, though, once you have the tests implemented. Building the tests is the hard part; comparing the results will be easy.

good luck

View solution in original post

DChiesa

I think your strategy for testing APIs hosted in Apigee Edge is not different than the appropriate strategy for testing APIs hosted anywhere else. I don't think there is special documentation in this regard. Just invoke the APIs using the appropriate dynamic request content.

As for test execution:

use a suitable warmup period. Caches get filled, pages get filled.
vary connection times and payload sizes to model expected real-world transactions
try to model the number of API keys or tokens - you cannot run an effective test using only a single token if your real-world client base will use 5000 concurrent tokens.
run small tests for at least 10 minutes
conduct many different trials
run at least one long soak test - 48 hours or so

I also don't have any guidance for you in comparing tests implemented in JMeter versus tests implemented in LoadRunner. That seems like it ought to be simple to do, though, once you have the tests implemented. Building the tests is the hard part; comparing the results will be easy.

good luck

anilsr

Just on a separate note, Few things related to perf testing,

Never do it on trial account, By T&C, you should not be doing same on Apigee Edge cloud trial account.
You can do same onprem without any issue on your environment other than production.
Don't do it on cloud with out performance pack, You might end up consuming your API Entitlements.
There are some add on packs for cloud if you intend to do performance testing.

Hope it helps someone looking for more details & license aspects.

-------------------------------

Anil Sagar

Learn Apigee Concepts in 4 Minutes HandsOn

optimism

Hi @sureshtirumalap

I would suggest that you may want to be careful setting up load test parameters.

You could end up using too much of your cloud resources, hence billing some unexpectedly high amounts of money.🤔