Load Runner - Ramp up and Ramp down the schedule
V
any formula/logic plz
While i apply how do i need to get formalize the scenario schedule
(includes Ramp up and ramp down) for my real time scenario?
Take an ex: Login-Logout a web application
Tool : Load runner-Controller
I Created a script to login-logout 25 Concurrent User
While I schedule it for a basic schedule it is taking approximately
12M:15secons for 25 users
For real time how do i do ramp up while login and after logout how do
i ramp down the time.
How do i schedule this.
Every time for different app's how do i handle this?
Please suggest me in this.
Floris Kraak
> How do i calculate the ramp up and ramp down times for the application?
> any formula/logic plz
>
Is there a formula/logic for how the real users use the application?
Then use that :p
A bit more seriously, the best thing to do would be to use production
measurements as an input for setting up your ramp up times.
If that's not available, there is not a lot of formality. Just a bunch
of semi-conflicting requirements:
1) You want your rampup to be realistic.
2) You don't want your test to take forever to rampup. So more than 2
hours to get to full load is not really productive (and probably not
realistic in the first place.)
3) You want to be able to pinpoint as precisely as possible where the
AUT gets overloaded. This cannot be done if your rampup is too fast.
Preferably it takes at least 15 minutes to go to full rampup so that
you will notice immediately when the response times start increasing,
rather than only noticing that you just killed the application.
Given the above a good rule of thumb is to keep rampup somewhere
between 15-90 minutes. Adjust to taste, based on your requirements and
previous sets of measurements.
Note that this does assume fairly large amounts of vusers - I usually
work with a few hundred at the least - so if you only need, say, 20,
feel free to make the rampup shorter ;-)
> While i apply how do i need to get formalize the scenario schedule
> (includes Ramp up and ramp down) for my real time scenario?
I'm not sure I understand the sentence above.
> Take an ex: Login-Logout a web application
> Tool : Load runner-Controller
>
> I Created a script to login-logout 25 Concurrent User
>
> While I schedule it for a basic schedule it is taking approximately
> 12M:15secons for 25 users
>
> For real time how do i do ramp up while login and after logout how do
> i ramp down the time.
>
Not sure what you mean with 'real-time'.
As for ramp down - personally I don't believe in it. Some people argue
it's good for detecting resource leaks but frankly I'd rather have
nightly endurancetests (soak testing) running to detect things like
that.
Do note that without rampdown your average transaction response times
may suddenly peak when the scenario stops. This usually is caused by
fast users stopping before slow users stop. Because some transactions
are still stalled waiting for response the vusers waiting for them
don't stop immediately. Users that are not stalled can keep going and
reach the end of their iteration sooner. The result of this is that
the long running transactions get reported later, at which point the
average will peak.
> How do i schedule this.
>
> Every time for different app's how do i handle this?
>
> Please suggest me in this.
>
Regards,
Floris
---
'What does it mean to say that one is 48% slower? That's like saying
that a squirrel is 48% juicier than an orange - maybe it's true, but
anybody who puts the two in a blender to compare them is kind of
sick.'
--- Linus Torvalds