Performance Testing best practice(s)
We use dynaTrace in conjunction with Silk Performer to place repeatable transactions and loads on the system under test - and also use the Silk Performer plugin from dynaTrace with Silk Performer.
Normally I am inclined to run transactions (via virtual users) in a series such as:
1, 2, 5, 10, 25, 50, 100, 200, 300, ... to end up with a set of data that can be plotted showing where the upper limit will in all likelihood be.
Does anyone else here have a better way of doing this? Thanks in advance,
Steven Shippee
Olympia, Washington
Hi Steven
I think you have the right approach. It is called an Increasing Load Test. Ramp up the load until the system breaks. Starting with dynaTrace 4.2 we have introduced a new Load Test Overview Dashboard that will make it easier to identify the breaking point per Timer Name. Here is the link to the documentation: Load Test Overview Dashboard
Andi
If you are interested in some more best practices on performance engineering you can also download a chapter from our free Enterprise Performance Book: Chapter 3: Performance Engineering
Andi
