It's hard to say without looking at the details, but one thought is that maybe when you ran the performance recording, cache was hit, and when you ran the tabjolt recording, cache was missed. To prove this out, you can add ?:refresh=yes&:record_performance=yes to the end of the url and then should bypass any use of cache.
Thank you for answering.
I tried the parameters I taught, but the results did not change.
Since the measurement range of the performance record and Tabjolt's response time are the following recognition, it is correct that the measured value of the performance record is longer than the response time of Tabjolt, but is the recognition correct?
- Send request from client
- Receive the request on the server side
- Server side processing (including query execution)
- Send response from the server side
- Receive a response from the client
· Measurement range of performance record
1 to 5
· Measurement range of Tabjolt
(According to https://en.scribd.com/document/359712202/Basics-of-Tabjolt-A-Tableau-Performance-Tool the response time is 'measured value of Tableau query')