- Behavioural changes as a result of unannounced internal release process. Go to bed, wake in morning to app serving 500s (happened twice)
- Design flaw that ran so deep they had to redesign the datastore, insisting on people start migrating before they even had migration tools ready. Prior to that, at least one outage event required running the Google equivalent of fsck and leaving "/lost+found" folders in everyone's datastore (WTF?!?!? Not even once, dude!)
- Latency that varies according to the phase of the moon, and you're billed for it anyway.
- Continually changing architectural story around apps. Last year: Memcache is cheap, free, and shared! This year: Dedicated memcache, only $66/gb/month! 2 years ago: elasticly spun up processes! Last year: dedicated hardware thread, only $100/thread/month!
- Let's not forget the wild pricing changes depending on how much the App Engine team had packed in their crackpipes the night before
- Service characteristics you won't see on any other platform (e.g. DB query latency). So regardless of abstraction layers, your app inevitably ends up designed for a single platform
It's a platform that succeeds only at exposing users with RAM-sized datasets to planetary scale problems, all the while charged handsomely for the privilege of the self-delusion that some edge was gained through all the suffering. Seriously fuck all that. I could turn this into an essay but why bother.
God forbid the bill you'd get for actually scaling to a size that started to require whatever GA2EE really is.
Even if App Engine isn't tremendously valuable now, it's a growth market, and easily monetizable. Google isn't going to burn a bunch of customers who could be - if they end up being the next Netflix - worth hundreds of thousands of dollars in recurring revenue.
I wish people would stop with the hyperbolic Google Reader stuff. Google's bread and butter are web ads, and they're a shrinking market. As Facebook can attest, mobile is hard to monetize, and it makes a lot of sense to focus on areas where people can just pay you directly. AWS has already validated this market, Google has the capacity to deliver, they've just fumbled the execution wildly.
That kind of cynical thought should be eliminated.
Not that many.
>and Google has probably invested tens of billions and thousands of engineers into GAE
If we're speculating, I'd say, far far less.
Except for the parts they use themselves too, anyway.
Stop wallowing over reader and move on.
"Google will announce if we intend to discontinue or make backwards incompatible changes to this API or Service. We will use commercially reasonable efforts to continue to operate that Service without these changes until the later of: (i) one year after the announcement or (ii) April 20, 2015"
So you have a year, at a minimum.
"Unless otherwise noted by Google, material changes to the Agreement will become effective 90 days after they are posted, except if the changes apply to new functionality in which case they will be effective immediately. If Customer does not agree to the revised Agreement, please stop using the Service."