When we have a new application, recording the application is a pretty tough task. We have tried multiple things. We do scripting or try to record with different settings and on different machines. We try to record multiple times, but we do not know why it is recording and why it is not recording. We do the same thing on different machines. It sometimes records, and at other times, it does not. That is one of the major concerns. Another issue is that while executing the load test, we recently faced an issue related to the mmdrv process. After the test is complete, from the LoadRunner Enterprise perspective, it does not show anything running, whereas, at the backend, we see that the load is getting generated from the load generator. We got to know that it is because of the rogue mmdrv process that is running even after the test is complete. We have raised this issue multiple times. The patches or solutions that they have given are not working. We have raised a ticket for this one as well. For each and every release, one of the criteria that we have is that it has to pass performance testing. It does not matter how big or small the changes are. Irrespective of the extent of changes, performance testing is required. We give a lot of value to performance testing. For us, performance testing is not only the test execution. We consider that as only 10% of the end-to-end performance testing. The analysis portion of it is where we want to spend 90% of our time. Our scripting effort is not too much, but it is still considerable. For a given application, if end-to-end testing takes 16 hours, then out of that, we need to spend at least four to five hours on scripting, customization, test data setup, etc. For the rest of the things, we have to spend about 12 hours. LoadRunner Enterprise can provide value by enabling us to do these analysis and performance engineering aspects within the tool. LoadRunner has a new feature in the cloud for single transaction analysis or single-user testing analysis. It shows all the insights in terms of the response and hot spots. If they can expand that to LoadRunner Enterprise as well, and not just to the cloud, it would be helpful. One more issue is that some of the features are enabled in certain LoadRunner versions. They are not available in other versions. To get a specific feature, the problem is that I either have to upgrade to a different tool or completely get rid of it and then use another tool. Instead of that, they can enable the features irrespective of the platform. Each and every tool that they are producing has its own advantages as well as improvement areas. For example, LoadRunner Professional has the advantage of using chaos engineering tools, which is not there in LoadRunner Enterprise. Irrespective of the tool version I am using, I should have the same advantage. I would like to see chaos engineering in LoadRunner Enterprise. Their support should be more efficient. I do not raise a ticket just to see the answers that are already available on the Internet.
