I have designed a case to wait for specific model name in GET response and then POST data. It is tested and is working fine when only 1 user and 1 loop are set. When the specific model name is found, then proceed to deploy the model.
JSON Extractor
$..valid_model_list[?(@.model_name =~ /.*Model-Test-${variable}-${counter}-v1-8k/i)]
While Controller
${__jexl3("${modelnameresult_matchNr}" != "1",)}
However, when I ramp up the test to use 2 users and 2 loops. The result isn’t as per expectation.
Question 1: Why JMeter isn’t running the request in sequence when multiple users and loop are set?
How do I design it in a better way to avoid false-positives in results?
Question 2: The GET (TrainedModelsList) constant timer is set to 1 minute and expectation is that it continues running till the end of the test. It seems it just run to a certain period and stopped. What type of timer should be used in this case?
JMeter deployed these model successfully
Model-Test-2022-02-28T09:46:40784-0001-v1-8k
Model-Test-2022-02-28T09:46:42785-0002-v1-8k
Model-Test-2022-02-28T09:46:44784-0003-v1-8k
Model-Test-2022-02-28T09:46:46783-0004-v1-8k
JMeter start deploy the model and fail although these models are not yet found:
Model-Test-2022-02-28T10:02:51745-0006-v1-8k
Model-Test-2022-02-28T10:18:18831-0007-v1-8k
Model-Test-2022-02-28T10:34:21111-0008-v1-8k
Model-Test-2022-02-28T10:50:24376-0009-v1-8k
Updated with view result notes where the problem occurred at second loop:
Source: Read More
