Slow redirects happening in gatling

Gatling (Open-Source)
1

Dear all,

I’ve got a scenario with a problem I cannot figure out why is happening.

I’ve got owasp zap as my proxy where I can see all incoming/outgoing requests/responses.
I’ve got a login process where there are 3 redirects happening before coming to end page.
In Zap each redirect only takes 50-80 ms, which is expected. But in Gatling one of the redirects takes 2.5 seconds, constantly!

And my code is fairly simple:

part of the gatling scenario where the redirects happen:
.group(“authentication”) {
exec(http(“/pkmslogin.form”)
.post(“/pkmslogin.form”)
.headers(form_urlencoded)
.param(“”“login-form-type”“”, “”“pwd”“”)
.param(“”“username”“”, “”“${username}”“”)
.param(“”“password”“”, “”“${password}”“”)
)

I can also see in my gatling logs that one of the redirects takes 2.5 seconds (Log is modified by taking away sensitive data):

POST:

2014-05-22 13:18:37,647 TRACE [GatlingSystem-akka.actor.default-dispatcher-4] c.e.e.g.h.a.GatlingAsyncHandlerActor -

Request:
/pkmslogin.form: OK
HTTP request:

POST https:///pkmslogin.form

…
…
REDIRECT 1:

2014-05-22 13:18:40,219 TRACE [GatlingSystem-akka.actor.default-dispatcher-7] c.e.e.g.h.a.GatlingAsyncHandlerActor -

Request:
/pkmslogin.form Redirect 1: OK

Why does the redirect happen so many seconds after the original post? Is there any way to debug this, or anyone can shed a light on this matter?
The big problem with this is that it tampers the statistics big time… :-/

2

I am using gatling 1.5.5 btw.

3

Hi,

Zap=???

https://www.owasp.org/index.php/OWASP_Zed_Attack_Proxy_Project

https://www.zapbi.com/default.aspx

There’s so many tools around that you can’t assume other people know the ones your using.

Does the problem occur under load, or even with a single user?
Do you play the same data with both tools?

Could you try upgrading to Gatling 2 snapshot?

Stéphane

4

Hi,

Zap = " I’ve got owasp zap as my proxy where I can see all incoming/outgoing requests/responses. "

The problem occurs even with a single user, and it’s always related to the first redirect happening.
Yes, all the calls made from gatling is going through the zap proxy, so it is the same data.

I will try upgrading to gatling 2 to see if there is any difference.

5

Hi,

Could you give Gatling 2 snapshot a try?
Does the slowness appear when setting Zap on Gatling, or directly?
Do you have a sample so we can reproduce, please?

Regards,

Stéphane