4 ms·
Our site was featured on CNN and Mashable this week. Here's what I learned.
- wesleyzhao 16y agoGreat takeaways Dan! Next time we create a fun app again, we will spend 72hrs awake tending to a whole other set of problems! Important thing is we'll enjoy the ride and keep learning. Love the post.
- dshipper 16y agoThanks! It's true but I guess it wouldn't be fun if it wasn't difficult. Let's keep this going.
- savramescu 16y agoNice article. I think that first step in getting a start-up is the hardest. I've been trying for some time, but I'm a programmer to the core. Never understood design. Whenever I found someone interested in something will happen for /sure/ and the whole thing goes down. I'm trying to compensate now by learning to do a basic design so I can jump-start something. I can always /pay/ someone for a design later if the website gets traction. /hat tip for the story
- dshipper 16y agoThanks for your thoughts. I really think a key to design is to keep track of websites that you like the look of and try to make something similar. Once you've mimicked enough designs you'll be able to do more original things. All it takes is enough time in Photoshop. Good luck!
- wacheena 16y agoGreat work. Can you expand on the FB oauth_id comment, how are you storing the ids now?
- dshipper 16y agoNow we store them as big ints. It has really helped us. What other details would be helpful to you? My email is dan at danshipper.com if you have more questions
- wesleyzhao 16y agoThis advice was thanks to the fabulous support at Rackspace (specifically Paul, ask for him). We had been fiddling with changing big queries to little queries, processing upfront vs behind the scenes, and nothing worked! All we did was add a couple intval() calls into our php script and changed the column to big int AND add an index and we had one 30s query go down to 0.6s. So: Use ints whenever possible. Index.
- danielharan 16y ago/me shakes head in disbelief "Index" is by far the big lesson here. Also, 0.6s is still way too high. What kind of funky join are you doing?
- wesleyzhao 16y agoWe actually have no joins haha. The support always asked if he had joins but no...I actually don't even know how to do joins. But I hear that Facebook highly encourages not doing joins anyways...so efficient learning?
- danielharan 16y agoWTF mate? 0.6s is WAY too high if you're not doing a join. Catch me on IRC if you'd like a hand.
- dotBen 16y agoWe originally had our site on a GoDaddy VPS server Developer friends don't let other developer friends use GoDaddy VPS/hosting.
- dshipper 16y agoAgreed. Awful experience we hope not to repeat.
- wesleyzhao 16y agoOur first big viral hit! We're learning!!
- underwater 16y agoFor anyone looking at his app be warned that it spams your Facebook wall without asking for permission.
- dshipper 16y agoWe are about to add something that states that on the home page. Sorry about that our intent isn't to spam just to spread the app.
- jayzee 16y agoAn intent to 'spread the app' without notification is spam
- dshipper 16y agoThat's true I guess I didn't express myself as clearly as I could have. What I meant to say is that our intent in posting was in no way malicious and we are going to add something that tells you we are posting to your wall. I agree that it's definitely a necessary addition so that our users know exactly what theyre getting
- laf2019 16y agoGood going guys! It's great that something you worked on at such a young age did well. While it is true that you should keep scaling in mind, dont waste too much time or money worrying about it before you need it. That's what makes cloud services awesome. By the way, how much did this end up costing you with bandwidth?
- dshipper 16y agoThanks for your thoughts I really appreciate it. And I think it's gonna end up being around $200 for the month.
- marcamillion 16y agoCongrats on your success...however, I have to point out something. When you said Rackspace, were you talking about Rackspace Managed Hosting (where you pay $1k/mo in a yearly contract) ? Or were you referring to Rackspace Cloud (where you pay per usage) ? Either would work from a technical point of view, but I think the former reinforces the notion of premature optimization. In your case, it was absolutely worth it - but how many people are going to decide to use InnoDB and memcached and get a trickle of traffic ? Probably the vast majority. I think you did the right thing, went with what was best for you at the time and migrated as needed. Although, perhaps you thought it was crazy stressful, imagine if you signed a $15K yearly contract and Mashable and CNN never covered you and you never got that huge influx of traffic. So yes, in your case, if you had moved to those specific infrastructure stuff before getting that large spike in traffic, it would have made your situation much more bearable. But for everyone else, that there is no guarantee that they will get the traffic - I would caution against preparing for the best case scenario (when unfortunately the likely result will be a modest amount of traffic). Just my $0.02.