Having been on the team that issued postmortems before, I can tell you that we said as little as possible in as vague a way as possible while meeting our minimum legal requirements. Actual Facebook customers (i.e. those who pay money to Facebook) will get a slightly more detailed release. But the whole goal is to give as little information as possible while appearing to be open. As an engineer that makes me growl, but that's how it is in this litigous world -- don't want to give someone a reason to sue.
These companies average users are highly technical developers, while facebooks users are from a much wider demographic.
It’s not really surprising to me that Facebook is writing comms that most users will understand right now, rather than publishing detailed post-mortems straight away. You have to speak the same language as your users initially in these comms.
Although I wouldn’t be surprised if we see a post-mortem in the days ahead, but Facebook probably will want to say why it happened (not just what happened, but why did it not get detected during testing, was the configuration change correct but there is an underlying bug on the routers etc) and what new mitigation’s will be put in place to stop it happening again, and these might not be known yet.
The thing is that a long term roadmap for a startup is 3 years. So you're going to look for things that are going to get you 3 years down the road cheaply, and then you still aren't that big, you have gigabytes of data, not hundreds of terabytes of data, so pivoting to another cloud provider isn't a big deal. Now, scale up to a company that has a monthly IT spend of 50 million dollars. You aren't moving that much infrastructure between cloud providers on a dime. Past a certain scale, there is a huge incentive to standardize on a vendor that has a 10 year time scale. We know AWS will be there in 10 years. Will GCP be there in 10 years?
I suspect the 50 mil / month guys will choose between AWS/Azure/GCP based on credits + discounts because all these clouds will be around 10 years from now. But it's true that I have no experience at that level (only from 0 to 9 mil / month) and the calculus could be completely different. I don't actually know if they could get greater commitments at that spend but it sounds likely. That's more than half a billion a year. Huge.
Perhaps someone who's made the decision on that can share.
They're already there on my part. Between their firing employees who organize or complain that their corporate slogan apparently now is "Be Evil", their chaotic product decision making, and their toxic internal politics, I have absolutely no interest in working for Google. Furthermore, I've noticed that the Google workforce seems to be becoming less.... Googly.... recently (disclaimer: I work down the street from the Googleplex and see many Googlers as they do their daily migrations). As in, they're now hiring a broader selection of people who aren't just the top 5% in class. What that tells me is that they're having more difficulty on the recruiting front and having to cast their net wider.
When I first wrote that sentence, I said "until people", but it occurred to me that being shunned by a small group of engineers probably barely registers on any metric that a large company would pay attention to. It wouldn't be actionable until the well had been running dry for a while.
I mean, shit, I work at a place where many of my favorite people have moved on to other things, few have been replaced, and yet we're only just having sincere conversations about how to fix our culture problems. That had to wait for one of the ringleaders to quit.
So far as I can tell from my minuscule keyhole, Sergey and Larry are the ringleaders. They'll never quit, and it'll take hitting rock bottom for them to have a change of heart. Not unlike Microsoft, after their stock began to tank.
The Fortune 500 made $1.1 trillion in corporate profits last year. The notion that they'll buy less stuff because shipping now costs $5 extra is ludicrous. Their problem is what to do with all that surplus cash now that they've bought down short-term Treasuries to basically 0% interest, a 1-3% increase in shipping costs wouldn't even appear on their "OMG!" radar.
