Measuring Logins with HTTP Monitors?

[AntonioSousa]

Traditionally, I measure synthetic logins with browser monitors. There is one case where this is too expensive for the client, and I'm considering trying to do it with HTTP Monitors. It seems feasible, but has anyone done it?

[kalle_lahtinen]

Hi,

I had the exact same situation earlier; there was no budget for large scale browser monitoring, but it was still kind of needed. All the apps used either NTLM or Kerberos. Back then NTLM support had just come out, but Kerberos not. Since then, it seems there's also support for Kerberos. At least the login via NTLM has worked perfectly. In case your application uses none of the available options i.e. Basic auth, NTLM or Kerberos, you'd need to handle the forms manually by customizing the HTTP request for the login. On a high level it looks doable IMO, but I don't have any personal experience about it. Of course one thing to remember is that HTTP checks don't execute any client-side JavaScript which a browser normally would. So if the login process requires that, it could be a problem.

[Bert_VANDEGAER_]

Hi there,

I'm currently trying the same thing, but for a login suing IIAMS so in other words with authentication ids and tokenids generated by HTTP POST commands.

It seems in theory very doable, but ... I'm currently blocked due to a limit of 1000 chars for using variables in pre/post execution scripts. As the generated authID is almost 2000 chars, I can't post the entire request body, so authentication fails.

Unless... someone knows how to maybe change this limit to +2000 chars???

Grtz,

Bert