Unanswered question

When a load test is run is the UI actually loaded or it records network requests and replay the same records?

The findings are quite interesting as it is using a notion of record replay, meaning it records network requests and replay the same records. The UI is not actually running during the replay, so the data measured do not represent a real E2E tests. Moreover as the network calls are recorded and replayed, all the data usually sent to gather UI performances information, are always identical (copy/pasted) preventing to use the actual performances data/boards


NeoLoad works at protocol level and gathers the server response time. This server response time should not be impacted by the UI rendering time. So the question is what you mean by "UI performances information"?


After some discussions with Team I have understood that they want to measure Time To Interactive (TTI) and not server response time.Which brings in another query:
Is it enough to measure TTI and not server response time for an application to perform under heavy load?