Why I will Likely Be Leaving GAE (30x increase in Price)

[Brandon Wirtz]

As many of you know, I am strong supporter of the GAE Community.  I was an Early Adopter, and I have defended many of the decisions Google has made.  I think the GAE Team is great, but I am loyal first and fore-most to my piggy bank.

 

I have 100s of small apps running CDN In A Box.  And they are all seeing similar price increase at about 15x through November, and then when the 50% discount on FrontEnd, goes away this app will go from $1.49 to $42.  One of the smaller clients will be going from .26 cents per week to 15.24 Pre-November, $28 Post.  That is over 100x increase in pricing.

 

When Google Announced the pricing change they assured me that the numbers I predicted in, increased cost would not be the 15x that I had predicted.  So I have waited patiently. But it appears that all of the small clients I have brought to GAE are going to be moving to Azure, unless something changes in the very, very near term.

 

My Business model is built around the people who want $20-100 a month services that I mark up 50-70%  If I sell a $20 plan that used to cost $1 and suddenly costs me $28 Things won’t quite work out.  I had planned for a 8-10x increase, but there is no head room on a 15-30x increase.

Resource Used Free Billable Charge CPU Time:  $0.10/CPU hour 13.93 6.50 7.43 $0.75 Bandwidth Out:  $0.12/GByte 4.28 1.00 3.28 $0.40 Bandwidth In:  $0.10/GByte 3.35 1.00 2.35 $0.24 Stored Data:  $0.005/GByte-day 0.00 1.00 0.00 $0.00 Recipients Emailed:  $0.10/1000 Emails 0.00 2.00 0.00 $0.00 High Replication Storage:  $0.008/GByte-day 12.55 0.50 12.05 $0.10 Backend Usage:  Prices $0.00 $0.72 $0.00 $0.00 Always On:  $0.30/Day No - - $0.00 Total: $1.49 Estimated Charges Under New Pricing The charges below are estimates of what you would be paying once App Engine's new pricing model goes live. The amounts shown below are for your information only, they are not being charged and therefore do not affect your balance. If you would like to optmize your application to reduce your costs in the future, make sure to read our Optimization Article. If you have any additional questions or concerns, please contact us at: appengine_up...@google.com. Frontend Instance Hour costs reflect a 50% price reduction active until November 20th, 2011. Resource Used Free Billable Charge Frontend Instance Hours:  $0.04/Hour 513.39 24.00 489.39 $19.58 Backend Instance Hours:  $0.08/Hour 0.00 9.00 0.00 $0.00 Datastore Storage:  $0.008/GByte-day 12.55 1.00 11.55 $0.10 Blobstore Storage:  $0.0057/GByte-day 0.00 5.00 0.00 $0.00 Datastore Writes:  $1.00/Million Ops 0.94 0.05 0.89 $0.89 Datastore Reads:  $0.70/Million Ops 0.18 0.05 0.13 $0.10 Small Datastore Operations:  $0.10/Million Ops 0.00 0.05 0.00 $0.00 Bandwidth In:  $0.10/GByte 3.35 1.00 2.35 $0.24 Bandwidth Out:  $0.15/GByte 4.28 1.00 3.28 $0.50 Emails:  $0.01/100 Messages 0.00 1.00 0.00 $0.00 XMPP Stanzas:  $0.01/1000 Stanzas 0.00 1.00 0.00 $0.00 Opened Channels:  $0.01/100 Opens 0.00 1.00 0.00 $0.00 Total*: (before clipping to daily budget) $21.41

 

Brandon Wirtz LockerGnome.com: Corporate VP Business Strategy BlackWaterOps: President / Lead Mercenary
Work: 510-992-6548 Toll Free: 866-400-4536 IM: dra...@gmail.com (Google Talk) Skype: drakegreene Lockergnome.com BlackWater Ops

 

 

 

[GAEfan]

Wow, Brandon. 14 CPU-hours and 4GB of data served per day? That is a
rather intensive app.

You have been a great contributor here. I hope you find a way to stay.

[keakon lolicon]

I think you should adjust Max Idle Instances and Min Pending Latency, they will much reduce the Frontend Instance Hours.

----------
keakon

My blog(Chinese): www.keakon.net

Blog source code: https://bitbucket.org/keakon/doodle/

--
You received this message because you are subscribed to the Google Groups "Google App Engine" group.
To post to this group, send email to google-a...@googlegroups.com.
To unsubscribe from this group, send email to google-appengi...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/google-appengine?hl=en.

[Brandon Wirtz]

Been there, Done That.

 

Most of the CPU time is from API calls to Fetch. I even use instance memory to cut down on Memcache, data-store and CPU time.  Trust me this apps is about as Lean as it can get.

[Greg]

We don't know yet how multi-threaded instances will affect all this -
if you are mainly IO bound (doing fetches) then you may get fairly
substantial savings.

Personally I wouldn't trust Azure with anything but test apps yet, and
I'm sure it has it's fair share of gotchas too. Appscale on AWS or
some other VPS would probably be my first port of call, but I'm
waiting to see how multi-threading pans out before getting too carried
away. Appengine has been very good to me so far, and I'm not giving up
on it just yet.

[Dani Shaulov]

I'm confused. How does it jump from 14 hours to 513 hours? And how does stored data jump from 0 to 13GB in datastore?

[Robert Kluin]

Hey Brandon,
Just wanted to note that I've heard the new pricing will bring a
128mb memory limit, so if you're doing a lot of caching in the local
instances it could start to be an issue too.

Robert

[Jon McAlister]

What's the projected bill like now that you've set max-idle-instances?

--

[Brandon Wirtz]

Max Idle instances seems to have very, very little effect. 

 

Scheduler seems to be broken, or optimized for different parameters.  From what I am seeing I think I went from being billed basically only for API calls, to being billed API calls plus the Time the Instance waits on those calls.

 

I also thing that Google Bot comes through spins up 100s of instances in 3 minutes and I pay for those for the next 15.

 

 

From: google-a...@googlegroups.com [mailto:google-a...@googlegroups.com] On Behalf Of Jon McAlister

Sent: Wednesday, September 07, 2011 1:56 PM
To: google-a...@googlegroups.com

[Jon McAlister]

I can't figure out which app-id you are referring to here. Mind letting me know what it is? It's hard to debug your case without it.

[Jon McAlister]

To follow up here, please see http://groups.google.com/group/google-appengine/browse_thread/thread/4ec68af1d6d5e998

It dives deeper into the data for this app, and is a good example of how max-idle-instances will reduce the bill, but depending on your request distribution, may end up hurting your serving latency and reliability. Each app will be affected differently, and developers should be setting max-idle-instances and max-pending-latency in accordance to the underlying tradeoff between cost and performance.

On Wed, Sep 7, 2011 at 5:41 PM, Robert Kluin <robert...@gmail.com> wrote:

On Wed, Sep 7, 2011 at 18:14, Brandon Wirtz <dra...@digerat.com> wrote:

Max Idle instances seems to have very, very little effect. 

 

Scheduler seems to be broken, or optimized for different parameters.  From what I am seeing I think I went from being billed basically only for API calls, to being billed API calls plus the Time the Instance waits on those calls.

I think that is exactly correct.