Long story short: we had to switch servers three times.
We currently run under a virtualized instance of Debian 6.0 (why: easier to maintain, backup, migrate, isolate, allows cloning and allows us to keep other dev instances). Until now everything has been running under Xen 3 and Debian, but once we upgraded to Debian 6.0 (both host and guest) things started to get bad.
You may know that the service has been unstable and all the time we had to work on it was invested on trying to troubleshoot and solve the issue (over 100 hours down in six months because of the still unknown issue). We first had problems with our provided that was limiting IO and network interrupts (and the whole system collapsed because of the high load queue). Then we switched to dedicated hardware with our own layer and got everything working. Then everything started to go down, sometimes once or twice a week but also a few times a day on some times.
After that, we switched hardware and everything looked good… For two days. We were believing that it had to do with the kernel and network layer as we were forwarding everything and doing routing with our own scripts between virtual network interfaces and the internet, and the NIC wasn't a good one. Switching hardware discarded that option, and now that we've discarded any external cause we can only believe that it has to do with the stock kernel and setup for Xen 3.1 available for Debian 6.0 stable.
Once again we're in newer hardware, changed everything and now we're running OpenVZ with yet-another-network-layer (that's the third time now). Hopefully we won't be down for much longer (except for 3-4 minutes at 0000Z some days of the week when we're doing full system backups).
So, yes: sorry for that. Hopefully we will never again be down for so much time or spend so much time without working on the important stuff on the web site. We've added some new shows, started the work for integration with thetvdb and looking forward to provide TPB-backed sources (point a tpb user as the source for a show, instead of using only eztv) for shows that aren't tracked by eztv.
Note: some data has been lost as the SQL datastore in the old server was corrupt in some parts, but not enough to recover from an older backup. For now you'll only notice that the schedule does not work. You can fix this deleting & readding the show, or just wait for a few days until we're fully running on thetvdb.