A few questions/pointers:
1) when you say number of users, is that what you're specifying in the --c parameter? That represents threads and doesn't equate 1:1 with users,. It's about 4-5 concurrent users per thread; so, for your capacity planning (300 concurrent users), you'd probably want a --c of 60 or 75 at the most.
2) you'll want to set your test duration to be at least 600 seconds to ensure likelihood of cache hits.
Please find my answer inline below .
when you say number of users, is that what you're specifying in the --c parameter?
Yes . We tried different combination testing but the main issue is if we run a test with 200 Plus concurrency count . Avg response time is increased to 3 minute .
We have a requirement wherein business will deploy the reports and 300 extra user would be added . Business want to make sure Tableau server will be able to handle the load if 300 users concurrently login and try interacting with the reports.
We have heard from Tableau vendor team saying 8 core can handle 112 concurrent load . We have 32 V CPU core licensed so want to know will my server be able to handle the load for 300 concurrent users.
If you're specifying 200 in the --c parameter that is equivalent to 800-1000 concurrent users, see my original reply. If you want to test for 300 concurrent users then you'd use --c=75 or so.
A quick question . Ive triggered a test run for 50 V users and can see avg response time being shown is 4+ minutes . I've used 5 reports in the test and when this report are opened manually will take 15 to 20 sec each .
So what exactly does AVG response time mean above for 4 + minutes . is it like for each user will take 4+ minutes when 200 concurrent session are opened (--c = 50) so concurrent session being opened will be 200.
Can you please clarify the above question.