It’s been a busy few days, sorry to everyone who’s contacted me and that I’ve not replied to.
I’m glad we’ve managed to do what a surprising number of “start-ups” fail at — starting. As you’d expect, traffic to our site spike after the TechCrunch article, and stayed pretty high for a day or two. I don’t know exactly how many uniques, but probably between 10 and 20 thousand, and we’re converting about 10% into users. We seem to be getting a second wind at the moment, as word spreads out.
We’re hosting on EC2, with a cool scaling agent / database caching proxy doo-dah to help us grow in the future. The guys that are developing the aforementioned agent / proxy doo-dah are actually staying with us in Pat’s house, which was really handy when it fell over in a heap at 4am on Sunday morning. Back up and running pretty quickly, although we’ve had brief outages pretty often as we’ve been patching like mad.
For those of you that have signed up, you may have noticed that search is a bit crap at the moment. I’ve spent the last couple of days wrestling with the python bindings for Lucene before facing up to the fact that hand-patched experimental code of alpha products is not good enough for production. So, just scrapped all that nonsense and tomorrow I’ll be moving over to use Solr via JSON. Hot replication and distributed indexing all for free like beer and speech.
