Yes, blame the operators. The operators designed their apps based on the original pricing incentive published by Google but got the rug pulled under them when Google hiked the price to completely change the pricing incentive and invalidate their design.
I got burnt by AppEng, too. Picking AppEng as a platform is one of my worst technical decisions.
"Google App Engine is free to use during the preview release, but the amount of computing resources any app can use is limited. In the future, developers will be able to purchase additional computing resources as needed, but Google App Engine will always be free to get started."
You got 2+ years of use for 'free' and now they decided to turn it into a supported business model and are asking you to pay for what you use. Seems reasonable to me.
I don't disagree that they fubar'd their original pricing release announcement and should have had multithreading Python 2.7 for those folks.
But they did listen to the (loud) feedback, made adjustments and even apologized (were you there at the ThirstyBear meetup where they bought us all beers?).
Not having to hire an IT staff or be woken up in the middle of the night when AWS decides to reboot the host and your servers go down is worth its weight in gold.
Yes, stupid of me to trust the marketing speak in that blog post. I've already said it was one of my worst technical decisions. I've advocated to clients to use AppEng and I looked real stupid when the projects failed because of operation cost hike.
Software are architected and designed with constraints in mind. Features are feasible or unfeasible because of these constraints. Cost is one of the big constraints. With the new cost structure, the apps have to be re-architected and redesigned, lots of things aren't possible.
I didn't go to the ThirstyBear meetup because I have given up on AppEng and moved on. I simply do not have trust for them to be a platform vendor.
If you correctly architected your application around the constraints of AppEngine, then it wouldn't be costing you a lot of money. That is the whole point... you know the constraints, so you architect wisely and you spend money where you need to. You decide which indexes you need and which ones you don't need. You decide how many reads and writes you are going to make. You figure this stuff out in advance and if you make a mistake, you refactor your code to improve things.
I'm not saying it is easy, it isn't. It takes a skilled engineer to learn this stuff and make it work. But, when it does work, it really does work.
It sounds like you failed to create an application that took all of that into account and of course you are going to look 'real stupid' when your clients figure out that you took shortcuts. There is no way that could be the fault of AppEngine.
I'm curious where you went after AppEngine. Are you hosting on AWS now? Heroku? EngineYard? Do you have the same reliability and scalability as what AppEngine provides? Maybe that isn't something that is beneficial to you and in that case, I can see how AppEngine is not your cup of tea.
For me, AppEngine is amazing. I love the fact that I'll never have to hire a sysadmin. I love not carrying a pager. I love knowing that when my site gets an asston of traffic, I won't ever have to think about making it scale. I love not having to worry if my database is on big enough hardware, replicated across data centers, backed up. I'll never have to think about whether or not my OS needs an upgrade to plug a security hole or ssh'ing into a server at 2am to figure some esoteric problem out. To me, all of these things are worth the 'cost' of AppEngine. I'd rather spend my time adding features than doing sysadmin.
At certain point, I am not sure whether you are trolling or not. All I see is you kept repeating the AppEng marketing nonsense.
A skilled engineer would know not all problems are the same and one platform can't solve all the problems. You don't know what product requirements I had and you assert it's my failing since I couldn't beat my product into AppEng's square, despite the AppEng's square turned into a circle.
Just because your niche app happens to fit into AppEng's mold doesn't everyone else can do the same.
We don't know your application, so of course we can't tell if you made poor design decisions or not. But have you considered that the poor match went both ways?
Google said that they adjusted their prices because people were getting wrongly incentivized to do things that were expensive at their end. And the way free market is set up, pricing is one of the ways that signals are sent to tell customers that they should make different optimization decisions --- including, perhaps, switching to a different technology which is a better match for their requirements.
Yes, that can be painful --- but it's the free market. You might as well complain that some silly folks moved to exburbs hours away from their work, and bought gas-guzzling SUV's, and then got upset when the price of gasoline went up to 3-4 dollars/gallon. Whose fault was that? OPEC? Or the consumer for choosing to live far away from work and to buy a car that had horrible mileage?
Valid points except "people getting wrongly incentivized." People were rightly incentizivized to optimize their apps under Google's pricing guideline. It was Google who got it wrong to misalign its infrastructure cost and pricing, and unable to bring down their cost over the years.
I don't think your gas example are relevant but if you want to stretch it, it just means Google like OPEC is not a trusted platform vendor.
I'm definitely not trolling. I'm trying to have a conversation and understanding of your rationale for why GAE didn't work for you. So far, you've left me totally confused.
I heartily agree with you about GAE being a niche. Not all apps (or developers) belong on GAE.
I ask again, if GAE didn't work for you, where are you hosted now?
Now look who is trolling. First you respond with the now deleted statement, which I'll post back here:
"ww520: You were not trying to have a conversation. You were trying to have a bragging session. I don't find it constructive to continue the "conversation."
Then you come back and respond again? I'm so confused. Anyway, I'll consider this thread over, unless you actually want to have a real grown up conversation.
It's funny to see a hotshot troll feinted confusion. It's even funnier to see a rude brat pretending to have a grown up conversation. If you want a grown up conversation, act like one first.
> (were you there at the ThirstyBear meetup where they bought us all beers?)
How is "must live in the Bay Area in order to receive friendly customer service" a reasonable hidden addition to the T&C of a supposedly global service?
I got burnt by AppEng, too. Picking AppEng as a platform is one of my worst technical decisions.