What do to when you aren’t sure what your servers can handle…

Crossing the line from an original idea over to real world performance struggles should be easier than it is. But, time and time again my observations suggest that 25-40% of available development bandwidth goes to maintaining and optimizing what you built and NOT to ongoing new feature development. Pages get slow. Users complain. Errors become more prevalent. Traffic spikes impact sales. It’s all a big struggle. For everyone.
I’ve been wondering how I would plan in advance for these challenges and that’s led me to reconsider how I would plow forward on a new idea. More details in this regard later because it impacts everything from design decisions to PR plans. Anyway, it goes without saying that your ability to conduct some manner of testing wherein each user is simulated by a separate thread with his own session information shall be important to understand how efficiently your app and hardware platform is.
I’ve been evaluating various tools for this purpose and am trying to collect information from others in this regard. The tests that I think are meaningful would include:
- Maximum number of users the webserver can accommodate before producing error messages.
- “Entire website performance at the normal (expected) load.
- Single URL tests of a webserver or web application to identify and discover elements that may be responsible for slower than expected performance.
- Testing for specific web pages which can be requested simultaneously without problems like database deadlocks.
Which tools do you use for NON Microsoft platforms? What other tests do you deem important? Let me know. I’m curious. Being curious is what we do here.
comments
One Response to “What do to when you aren’t sure what your servers can handle…”
Leave a Reply
Interesting you posted this as FD’s servers got slammed big time today with a DDOS (yes, I decided to help out - mostly resolved now!)