Loadtest with one master and multiple slaves is failing at higher rates than expected
Here is my situation.
I have a load test with a scenario. its runs ok on the master machine with acceptable error rate of say under 1% and acceptable limits for through put.
Now when I add another slave machine and run the same test and scenario, my error rate jumps to say 35%. Has anyone experienced this? What could be the reasons we are seeing this?
In my case both slave and master are on the same host but perhaps running on different ports like below:
Comments
-
My first thought is that you are pushing more load on your application and it can't handle the load. Could that be the case?
0 -
@jakuniak, No that is not the case. Our customer is doing a 4x load test from their app (integrating with our APIs) and it is flying through with reasonable SLA numbers. we are only attempting one half of our customer's test (we are not overlapping tests)
0 -
I've since tried adding another machine, this one being remote. I'm running into the same problem..Appreciate any help or insight anyone can provide.
here are updated screenshots
0 -
@jakubiak, let me know if you need anything else to troubleshoot. I see a lot of errors w.r.t to webtool not sending the correct requests to the API endpoints. Another observation is the test progressively fails from top to bottom (assuming have 10 tests, 1-6 run ok, but 7-10 keep failing). And not all tests fail for test cases 7-10. Only some do.
Any help is appreciated.
0