Twitter Status (well, the fake one anyways)

Same story, different month

Tue, 18 Nov 2008 11:30:47 -0500

Its like you can only pose for so many front-page exposes on your company before you have to start doing some work. My favorite was when <a href=”http://twitter.com/ev”>ev</a> was talking about how we’ve been spending the past few months building new features for Twitter. Ha! I love it. Writing features?! Fuck that. We’re taking features away my silly little friends.

Which brings us to the last week or so. You know, we could get up here and tell you that our SMS aggregator isn’t working so many times before you guys figure it out. Because let’s be honest; SMS is not hard. The reality is we’re bleeding cash and we figured it would make sense to just sort of randomly drop 40% of our SMS messages on the floor. Its like roulette now. Only better!! Now, we’re the geniuses for cutting costs by a full 15% (yeah, you do the math suckers) and we have the means to wratchet that up or down. Its just you silly users that suffer.

And then there’s this morning. “Database upgrade”. Haha. OMG. Who comes up with this shit?! Oh wait, we do. Ha! Because that makes sense. Let’s do an upgrade on a Tuesday morning PST. Totally makes sense. Retards.

Here’s the deal; as you all know Twitter is a really</strong, really hard problem. Like harder than porn. I mean seriously. 140 chars to like almost a million people?! Yeah. Think about it. 1 million people.

The database is fine; we’re just waiting for our sysadmins to get in (they usually get in at 10:30am or so - its the Valley baby) to “reset” the servers. In the mean time its motherfuckin’ mimosa’s and Krispy Kremes.

New features, fun but don't expect change

Fri, 26 Sep 2008 12:40:19 -0400

Sure. We can’t really keep Twitter up. We’re introducing new bugs all the time. Whatever. Get over it. Its not like IM and Track have been off since May. Oh wait. They have.

In any case, we’ve been hard at work on a bunch of features you could care less about. You’ll notice a sexy new “Election ‘08” banner across the top of the site now. It takes you to a version of Summize that doesn’t really do anything you couldn’t already have done before.

Now, we did just realize something after we launched this and we think its at the root of all of this whining/bitching/moaning going on. You people are all Obamaniacs. You all want change. You all demand better. Well, sorry folks. We’re runnin’ this like Bush and McCain; a whole lot of the same crap with lots of hand waving.

Logging in

Thu, 25 Sep 2008 23:36:38 -0400

WTF?! You want to login?! Jez-sus people. Get over yerselves. Don’t you people understand we’re in the middle of a financial crisis?!

We’re blaming that bitch Katie Couric for our latest outage; how could she ask the defender our our Northern Frontier about her experience in foreign affairs. But we digress.

Missing Tweets

Tue, 23 Sep 2008 18:48:00 -0400

We find it hilariously ironic that you people think you own your tweets. “Where are my tweets?!” you keep bitching. Read the TOS you morons; once you tweet it, we can do whatever the hell we want with it. Including lose it. Or put it in someone elses profile. Or reverse the characters or munge the punctuation. Why? Because we can.

In all actuality, we had some issues with our cashing servers. Although this goes against traditional thought on the matter, once you take a big funding round, you can in fact still have cashing issues. In our case, the cashing has manifested itself as a whole bunch of cocktails and a mega-hangover. Need we say more?

Can you guess what we’re going to say? We’re still working to fix it.

Suckers.

Unplanned outage

Tue, 29 Jul 2008 12:18:10 -0400

Its a case of the Mondays but on a Tuesday. WTF?!

We’re having some “issues” with our “database” so we had to go “offline”.

“Issues” you ask? Well. Its rather complicated. You see. Our database cheated on us. I know, I know. It was embarrassed that we always point the fingers at it and not at ourselves. It was crying out for attention and we simply wouldn’t listen. That and we don’t cuddle after sex.

We’ve paged our therapist in for a quick couples (group?) session and hope to have the “issues” resolved soon. Like any issues like this, the odds are there will be lingering effects for weeks to come. Get the fuck over it.

Followers and Followings == FAIL

Thu, 24 Jul 2008 20:45:05 -0400

WTF people?! Okay, okay. We know something is up and we’re working on fixing so STFU. Geez. You’d think you people depend on this service or something.

Here’s the deal. We think we know what it is. We sent a bunch of our engineers on a mad dash up to Portland, OR for OSCON because we’d heard (via our latest acquistion Summize no less) that there were a shit ton of engineers up there with all kinds of ideas on how to “do Twitter right”.

Guess what? You’re all fucking wrong. We tried doing that C++ application with a bunch of linked lists and either we don’t know what the fuck we’re doing (hint: not likely) or it turns out Twitter is an even harder problem than we thought. Can you believe that?!

Amazingly there are some pretty smart people here but it turns out not smart enough to fix Twitter. Google people offered up suggestions. FAIL. Yahoo. FAIL. Pony-tailed, bearded guy using obscure databases? Double FAIL. As such, we’ve proved that there are no smart people that use or develop open source.

We’re headed to Redmond tonight to hopefully get some knowledge from our MS peeps. Weez out.

Summize acquisition complete, let the sucking begin!

Sun, 20 Jul 2008 12:59:25 -0400

We have finally subsumed Summize and we’re doing everything we can to fuck up the site. To that point, we’re removing several really handy features as well as generally slowing it down.

Honestly we’re glad we acquired these guys; they were really, really starting to make us look bad. As an added bonus, its a lot easier for us to find Twitter haters again and we’re continuing to add to our “list”. That’s right, keep fuckin’ with us. We know who you are.

Interviews with mean people

Sat, 12 Jul 2008 12:40:03 -0400

This has nothing to do with the status of the Twitter service. However, it indicative of the grief we pour souls have to endure every, single day. We work hard to provide less than mediocre service. You people should be happy with us.

The following are a few excerpts from an interview we did with “Asshole Interviewer” (hereto referred to as “AI”). We could have let them publish this on their own site but we snuck in the requirement that the interview could only be 140 characters long in the agreement and the guy (Err. Maybe its a girl?) lost his shit. Pretty funny actually:

AI: Thanks for doing this interview with me today gang. I really love the Twitter service but like many users, I’ve been frustrated with the recent outages and abundance of Fail Whales. What have you all been doing to help fix the situation?

Us: Fuck you. Nice t-shirt. You wore a fucking Fail Whale (tm) t-shirt to this interview?!. Guess what? We’re trademarking Fail Whale (tm) bitch. Pay up. Oh, and you call this a “situation”?! WTF. This is a free service for like maybe a million people. Maybe. Oh yeah, make sure you make “free” and “maybe” in that last sentence bold.

As for what we’ve been doing, well, that’s none of your damn business. We’ve been working hard to “increase capacity” and have added many “slave databases” as well as tweaking our “load balancers”. BTW - make sure to put the buzz words in quotes; we’re not really sure what they all mean but all the blogs we read keep saying we need to do those things. Wait, why are you writing that down? That’s off the record. WTF. Knock it off.

AI: Sure. I’ll just leave that out. No problem.

Us: Great. Or else its sleep 2 for you bitch. Don’t mess with Texas.

AI: Alrighty then. You guys talk a lot about Twitter being a really hard problem. What are some other hard problems on the Internet that need tackling?

Us: Sweet. That’s like underhand softball. You know what’s hard? Search. Now, imagine doing search for the entire Internet. Fucking impossible, right? I put that in a category right below what we’re doing. Technically the means to search the Internet is out there, but nobody has …

AI: Um. Ahem. What about Google? Don’t they search the Internet and make it look relatively simple with little or no downtime?

Us: Fucking A man (Editor: or maybe a woman). Shit. That’s a good point. Hmm. I guess they do make it look pretty easy. Okay. Okay. Wait. I’ve got one. Pictures. Serving pictures for like hundreds of people a day. Fucking impossible and I’d like to point out, nobody has cracked that nut. Again, it seems like it should be easy to do but find me one example of somebody doing it well and making it look simple.

AI: Flickr, Photobucket, Shutterfly, Picasso and SmugMug just to name a few.

Us: Shit. Damnit it. I said one. Geez. You’re kind of a know-it-all asshole interviewer, you know that?

AI: Hey, you asked.

Us: Guess what? Sleep 2 motherfucker.

AI: What?!

Us: Don’t worry. I’m sure you’ll figure it out. Oh wait, I’ve got one! Video. Nobody has …

AI: YouTube, Vimeo, Blip.tv, Seesmic, Flickr (again). Shall I go on?

Us: Fuck you. This “interview” is over. Make sure you put interview in quotes in that last sentence. This does not count.

You know we really, really try here at Twitter. Serving 140 characters to almost a million people is really, really hard. Harder than search for the Internet, photo sharing or video serving. Way harder.

Stay tuned for our next couple of updates where we’ll be describing the how’s and what’s of scaling Twitter.

Check us out, bitch

Fri, 11 Jul 2008 20:22:34 -0400

Yeah, we’re a little lippy today. Why you ask? Well, it has nothing to do with the fact that like half the company is at FooCamp. You can’t buy that kind of cool. No siree. Two words: up time. Fuck. That’s like one word. Whatever, look at this graph suckers!:

Now, I know what you’re thinking. How does compare to someone like, say, Google? It doesn’t. Google is way better. But that’s not the point. Google is lame. I mean seriously. Anybody can search the entire internet. We’re doing like instantaneous updates for the coolest people in the world (hint: if the service goes down over the weekend FooCampers will be sad).

So yeah. The site is up but it still takes over a second and a half to load. Cause for celebration? I think so.

Silly identi.ca, you're no match for our service foo

Thu, 03 Jul 2008 11:53:25 -0400

Awwww. How cute. Somebody thinks they can launch an “open” microblogging service. Fat chance. How can you compete with our rabid user base?! Its impossible. We beat our users senseless with outages, disabling features and general malaise and yet they still return day-after-day. You have a lot to live up to my Canadian friends.

It appears this morning that they’re having some issues as identi.ca isn’t that any more. Its pointing at hewitt.controlezvous.ca. Now, we don’t know about you all, but we call that a bait-and-switch (and hewitt.controlezvous.ca just isn’t all that catchy TBH).

Our favorite part is the fact that they use OpenID. C’mon. Seriously?! People actually use OpenID?! That’s cute. In a sad, hurt, puppy dog kind of a way.

So take that you lackeys to the North! Running a microblogging service is hard. Trying to defeat us is futile. Being open is silly.

We don’t care what anybody thinks. Mostly.

Quick "Upgrades"

Wed, 02 Jul 2008 12:49:00 -0400

You’re totally going to laugh about this one. Somebody tripped over the power cable in our new offices and through some kind of “Butterfly Effect” kind of thing, the entire website went down. Go figure.

In the mean time, we’re all just hanging out in the new office (those of us that make it in here before 10am that is) waiting for services to come back up. ETA is 10 minutes right now but I’m guessing it’ll take more like an hour or two. These things never go as planned (see last 2 years of Twitter as reference to the fact).

That’s funny, right?

PS - this does not count against our Pingdom numbers because technically the site is still up and serving fail whales. Uptime. Its all about uptime people.

Update: (10:10a) And we’re back!!! (in our best Jimmy Fallon). See, we jinxed ourselves into getting back up quickly!

"Fuck You Twitter"

Wed, 02 Jul 2008 10:24:00 -0400

You know what? Just stop it. My lord. Here’s the deal: we don’t know what we’re doing, okay? We’re just getting killed over here and all of you just sit there and write long diatribes about how scaling Twitter is trivial or hard.

After today we’ve had it up to about here with you people. Err users. Customers? WTF. Well, we’ve been loving Summize as of late (how do they do search so well?! - we should just hire those guys) and somebody jokingly put “fuck you twitter” into the search box. You know what came up? Well, you can look for yourself.

We know who said that. Guess what we did today? For every motherfucker that decided to talk shit about us in the last three days (you’d be amazed at how many people that actually is) we’re adding a “sleep 1” to every action you do on the website. Don’t like it now? Fine. Fuck you. Shut your cake holes and we might not crank it up to 2.

Tuesday morning randomness

Tue, 01 Jul 2008 10:36:00 -0400

We’d love to tell you that we know what happened this morning but we just plain don’t. This computer stuff is hard.

What we do know is that certain users are only seeing limited updates. Here’s what we’re going to do to try and track down the problem:

Do a little dance.
Make a little love.
Get down tonight.

Oh wait. That’s not going to do it. Shit. Okay, seriously. Let’s get serious for a minute. Lord knows you people will complain if we don’t act all professional. (Bonus for us: at least you can’t complain to us on Twitter when its down)

Okay, seriously. Here we go. This is what we’re going to do. Really. This is it:

Spend the day watching completely realistic movies like “Sneakers” and “Die Hard 3” to learn as much as we can about computing. They make it look so easy in the movies.
Check, re-check and triple-check the “database” and “load balancer” (honestly, we’re not even sure which is which at this point).
Compose a lengthy status update on status.twitter.com … doh … wait … already doing that.
Update the Twitter Wikipedia entry to reflect on our catastrophic failure to provide any service what-so-ever.
Contact the Guiness Book of World Records to inquire about “Worst System Engineering Ever” record. We’re pretty sure we’ve got a lock on it.
Cross our fingers.

And that’s brings us to now. So, yeah. Ummm. Happy 4th of July?!

Update: My lord people. Yes, SMS is down. Yes, the site is slow. You know what? All you people that are complaining? Why not just switch over to Pownce. I hear they can keep a service up. PS - we still kick their ass! Take that Leah!

Monday afternoon downtime

Mon, 30 Jun 2008 16:19:00 -0400

Are you sensing a theme here? That’s right. We’re hating life again.

One thing that would really help us right now is if folks would just login to Twitter and then refresh your page a few times. Really fast. Not.

We’re seeing slowness again this afternoon that we’ve traced back to our ineptitude at running a “major” Internet service. I put “major” in quotes there because when you look at the numbers, we really don’t have that many users. Its like 1.5 million maybe. So we should qualify that original statement; we’re seeing slowness again this afternoon that we’ve traced back to our ineptitude in running an Internet service.

Now, while we’re busy posting these status messages we could be working on our golf game, filing our (stupidly late) taxes or even fixing Twitter. But nooooooooo. You people have to have updates. Constant updates. If only someone would build a service that would notify you and your friends quickly via the web, mobile, etc when things are happening. Now that would be awesome.

Update: 1:46pm - headed down to the data center and we pressed the “Turbo” button on all of the servers. Hoping that solves this afternoons slowness.

Monday morning downtime

Mon, 30 Jun 2008 12:14:00 -0400

We effectively have two catch alls when it comes to problems here at Twitter; database and load balancer. This morning, our “load balancer” was “overloaded” and we had a few “hung” machines. Honestly, we’re amazed the Internet stays up as it does considering we can’t get a Linux box to function for more than 8 hours without it seizing up. Ya gotta wonder how Google does it sometimes.

Our best guess is that the Intertubes we have connected to the load balancer were flooded with all kinds of pornography from the weekend. Its taken us most of the morning to flush those tubes out and we’re still at a loss as to why the Internets are so slow. On the line with our co-lo now.

Update: 11:35am - phew, we’re still “working through this problem”. Ha. Just kidding. Its been a Mimosa kind of a morning. Please bear with us.

Saturday night downtime

Sun, 29 Jun 2008 12:40:00 -0400

Fuckin’ A people. We had another “database problem”. This is kind of a catch all for us now. This time it was because somebody thought they were in the right terminal window when they typed “kill -9 -1” to shutdown their machine. Seriously? I know I hate typing “shutdown” too but pa-leeze people. The ensuing chaos and subsequent downtime led to a promotion for said individual.

Here’s to hoping we can get our shit together for the 4th of July weekend.

Outage today between 2:30pm and whenever

Sat, 28 Jun 2008 17:23:11 -0400

@al3x met this chic last night that said she could “tweak the crystals” enough to make the site faster. Whatever, we’re endulging him/her. She’s got the con at 2:30pm and we’re hoping she can figure this shit out. What a mess.

Hope you didn’t need the site this afternoon. Here’s to crossing your fingers!!

Good Morning Twitterverse!

Sat, 28 Jun 2008 12:22:00 -0400

You know what we did last night to make Twitter better? Nothing!! That’s right. Between the late nights of drinking, all of our adoring fans and the rest of you suckers that put up with our abuse, we decided to take it easy last night. Its been a long week for us!

This weekend we might bring back replies. If we feel like it. We were kind of pissed at how all of you have been dissing us publicly so don’t push us too far. We’re fragile.

On the plus side, since we turned of @replies, uptime is through the roof!

PS - if anybody has any ideas on how we can actually monetize this sucker, we’re all ears. Just kidding; we don’t really care what you think. Hahaha!

Turning @replies back off

Fri, 27 Jun 2008 23:37:00 -0400

Just stop it you smarmy bastards. Putting the timestamps in the posts like @bradfitz. You should be ashamed of yourselves. We’re slaving away over here trying to make things better and all you people do is pout about it.

You know what? We’re just gonna leave the @replies off while we “work on the stability of this feature”. Keep messing with us. Just keep it up and we’ll make the website into a bunch of fucking ASCII art.

Friday Afternoon: one more thing

Fri, 27 Jun 2008 19:26:30 -0400

Okay, so the service is stabilizing. We think this has to do with the rapidly approaching Happy Hour across the West Coast. Or people are finally pissed off at us (just kidding; you suckers will never leave us no matter how we abuse you!).

Oh, and to the guy at the coffee place: no, Twitter is not like third-world power grids. We stay up. Our uptime is stupidly high right now. You can’t complain with anything less than 99.9%. And yes, serving up the Fail Whale counts as uptime. Geez.