jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Richard Gaywood <richardgayw...@gmail.com>
Subject How does throughput calculation work from a JTL file?
Date Wed, 09 Nov 2005 09:47:06 GMT
I'm rather confused... I've been collating some load test results this
morning and I don't understand what JMeter is telling me.

I've loaded my JTL files into the Aggregate Report element and I am told
that the three elements of my test (three different SQL queries in three
ThreadGroups) did the following numbers of samples and the following
throughputs:

type1: 5169 searches, 348/min
type2: 11917 searches, 804/min
type3: 2119 searches, 144/min

Thing is, I ran the test for five minutes (and the times on the log files
confirms this) so I can't see any way in which the througput figures are
right. Surely, for example, (11917/300)*60 = 2384 searches/min?

Opening the XML log file my SimpleFileWriter generated, I have the following
timestamps for the first and last entries in the file:
1131454087894
1131526605450

(all three search types ran throughout the test, so these approximate
numbers hold across type1, type2, and type3).

The difference between these two is 72,517,556 and I can't make that make
sense in any unit of time I'm familiar with. Can anyone confirm what format
the timeStamp element of the file output is in? If I could make sense of
that, perhaps I could figure out how JMeter is calculating the throughput
figures.

(I'd like to just turn to the source code, but unfortunately my time is very
short right now. However, I am hoping to contribute some code to the JMeter
project so I promise I'm not using your time up for nothing!)
 53517433

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message