[TIP] cloning http traffic

Kumar McMillan kumar.mcmillan at gmail.com
Thu Apr 3 07:58:06 PDT 2008

On Wed, Apr 2, 2008 at 10:41 PM, Titus Brown <titus at caltech.edu> wrote:
> On Wed, Apr 02, 2008 at 07:48:19PM -0700, Joseph Heck wrote:
>  -> Anyone have any tools or plugins that they use to "clone" http traffic?
>  See scotch, http://darcs.idyll.org/~t/projects/scotch/doc/
>  Should be easy to modify to have it do this, by hooking the recorder and
>  the playback directly to each other.

As Titus and I have discussed, there is a commercial tool called
TeaLeaf that also clones http traffic:

You configure it with your session keys and it sniffs the logs,
request / response bodies, and caches your resources daily.  With this
it lets you "play back" exactly what a user saw, what form values she
submitted, what error or success messages she saw, everything.  This
even works for dynamic sites and I haven't seen it impact performance.
 Actually, I'm not even sure that it sniffs responses, it might just
sniff the logs.  I've seen it used to submit more meaningful bug
reports and to find bugs that you didn't know existed or explain
definitively why orders aren't coming through, or whatever.  Where
error logs can be useful they don't always tell you what you need to
know.  Actually going back and *watching* a user's experience is
incredibly helpful.

It would be really really cool if someone used scotch to do something
like this so that its features could grow organically in the open
source community ;)  I think it would be a huge success.  OTOH, this
product is well worth the money.


More information about the testing-in-python mailing list