I think there are two answers to the questions: - My answer. The bottleneck is always the database. - Some people have had problems with dropped request. We have not got the bottom of what causes it. I cannot reproduce the problem with my setup.
In my experience for an average database driven application, a http request takes 20-40ms. This means, assuming a single server, 25 requests/seconds. That means your 500 users should not request more than one dynamic page every 20 seconds. Static pages (images, media) can be served directly by the server and do not count. This is ideal. There are things that can go bad (not web2py specific), for example: database needs to rollback some transactions and takes longer than usual, network is slow, connections take long time, and there is only so main the server (apache) can handle, depending on configuration. Massimo On 9 Lug, 22:32, VP <vtp2...@gmail.com> wrote: > hi everyone, > I've been looking at web2py for a little while now, and am still > learning. I'm thinking about using web2py to develop a simple social > network app. My question is how many users can this app handle > concurrently? Is web2py the bottleneck here? > > I am aim to get up to 500 users being active at the same time. Is > web2py appropriate for this task? If so, can somoeone give me some > pointers on how to set it up? > > Thanks > VP
