- ab - Good benchmarking tool
- twill - Awsome web testing tool
- jmeter - Slow Java Swing but probably good to playaround with
- httperf - Don't remember that much about this tool but it is pretty cool
If I could mimic user traffic it would be nice to have something that would be able to help me identify the little nasties I have created. Maybe something tied into twill would allow you to mimic traffic plus using the profile be able to point out the bottlenecks.
OK, the truth is that most of the nasties I know where they are. I am just to lazy to fix them. I shouldn't fix some becuase they are not heavily used features. I guess if the site gets slammed I spend all night at the office. I am sure this is some kind of programming design XP or Botchy, I like to call it lazy first.
I think that if TG had a tool to help find bottlenecks in TG then business logic coders like me could use it too.
On 3/23/06, ajones <
[EMAIL PROTECTED]> wrote:
It's that time again. I hope to come up with some thought provoking
open questions that will get new ideas out there, or at least provide
some intellectual relief from the more mundane project concerns(read:
probably more important). Anyways, on with the show.
We all know having a wildly successful website would be awesome. What
could be better, you toss your code out into the ether and find that
the clamoring hordes love it, cherish it, and all want to be one with
it 24x7. Everything is awesome, until your server catches fire and
melts into a puddle on the floor. Not quite so awesome, now you have
angry hordes and a big mess.
Obviously there are ways to prevent this, caching, replication, etc,
but how would you know when to implement them? Wouldn't it be nice to
have a testing system that can tell you when you max out on
database/request/processing load, which will hit the wall first, etc.
We run unit tests to make sure our code works as designed, what about
stress tests to figure out where the breaking points are?
What do you do to check the performance of your code? How do you
isolate problems that would only occur during high utilization. Should
turbogears include something to help you break down where your code is
wasting time instead of fulfilling requests? How would it work?
I could see modifying unit tests to add some kind of performance
metric. It would be fairly easy to hijack the existing test functions
and have them report how long everything took. That would be a decent
start. Where should it stop though? How do you load test a server
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "TurboGears" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at http://groups.google.com/group/turbogears
-~----------~----~----~----~------~----~------~--~---
