I manage a site for a Catholic school and parish.  We host it with PowerDNN, have only a couple of custom modules (Ventrian Simple Gallery and a couple simple ones I developed myself), and generally keep up to date on releases (running 6.01 now).

We experience a very frustrating and seemingly random occurrence of the display of the generic DNN "The site is unavailable" page about every 2 months or so.  There is never any correlation to any configuration, module, upgrade, etc. changes to the site, and there are never any errors or exceptions in the DNN Event Log indicating what the cause might have been.  When I notice this or someone complains to me that the site is down, I log into the PowerDNN hosting control panel and recycle the app pool to resolve it.

PowerDNN seems to be unable to help or resolve this.  The only explanation I have received is that this sometimes happens after Microsoft patches are applied and all of the servers re-boot, and if too many sites are starting up at the same time and trying to reconnect to their database server all at once, some may timeout and get into this "Site Unavailable" state.  They have suggested I subscribe to a site monitoring service so I can be notified when the site is displaying the "site unavailable" page and then go in and do the manual recycle of the app pool.

This situation is pretty unacceptable.  I am wondering if the solution is to find a new host, or if this is a DNN problem that would persist if I changed hosts.

Are there other places I should look for clues as to why this is happening?  

Are there other potential workarounds or solutions?

Well yes we have a couple of custom modules, a few I wrote and a couple of other ones like the Ventrian Simple Gallery. The site runs fine for months, going through many routine app pool recycling events as most sites do on IIS. Then wham-o, something happens in the environment (the windows patch/re-boot cycle explanation makes sense to me) that makes DNN abort its normal startup routine.

It is my experience that when anything goes awry at key events in the app startup, DNN just gives up and throws that "site is temporarily unavailable" page up and never tries again. I have most frequently seen this when version upgrades go off the rails, but that is in a controlled dev/sandbox environment where I am watching every step, researching and implementing fixes along the way to get through the install or upgrade.

These random production failures are much less tolerable, particularly since half the time it is someone else noticing them first, and since the 100% surefire fix is a simple recycling of the app pool.

Is there somewhere besides the Event Log where I should look for any telltale clues as to why DNN decides to give up its normal startup and throw the "unavailable" page up?