I maintain a Sitefinity 5 site that is fairly heavily customized, which I think plays a part in my problem, that takes upwards of 10-20+ minutes to become usable again after an app pool recycle is initiated. Sometimes I'll even have to kill the w3wp process for that specific app pool.
I'm wondering if there isn't a setting somewhere on the site/app pool that is causing problems.
So, I decided to create a new site, same codebase (different folder) with a non-routing hostname so only I can hit it, and it loads/recycles relatively quickly (1 - 2 minutes).
Does amount of traffic play into how quickly an app pool recycles?
Any other ideas?
I think the amount of traffic plays here, because once the app pool restarts, all cache is cleared and if there are many requests at the same time - they will all try to fetch the data from the DB because the cache is still empty.
In the recent version (7.3) there will be a switch that will control this. It will work in a way that the first request will enter a lock section and only that request will be able to get the data from the DB, then it will populate the cache and release the lock.
The other queued requests will then read from the cache.
More details are here: http://www.sitefinity.com/blogs/boyan-barnevs-blog-/2014/11/28/improve-the-application-start-up-and-the-effects-of-page-templates-changes-for-websites-with-heavy-load