Google App Engine - Region

[Hossein Amerkashi]

Hello,

I am using GAE and have deployed an app. The performance seems to be pretty good for USA east-coast. I'm getting complaints from UK users that performance, during certain times, is very slow. How do I find out what region the servers are and how can I upgrade for multiple regions?

THANKS

[Kenworth (Google Cloud Platform)]

On Sunday, September 24, 2017 at 11:04:16 AM UTC-4, Hossein Amerkashi wrote:

Hello,

I am using GAE and have deployed an app. The performance seems to be pretty good for USA east-coast. I'm getting complaints from UK users that performance, during certain times, is very slow. How do I find out what region the servers are

Console > App Engine > Dashboard > Region is located on the upper-right below your domain name.

 

and how can I upgrade for multiple regions?

App Engine is automatically scaled across multiple regions, but those regions are defined by Google. You'd have to set up another project to handle traffic in the UK, for example. You can visit this related thread that discusses possible workarounds for multiple regions.

[Hossein Amerkashi]

@Kenworth, So we have to deploy/manage another instance of project for Europe region??

Where does the loadbalancing come in play? Is it possible to somehow setup loadbalance which directs traffic to proper region?

Thanks

[Jordan (Cloud Platform Support)]

Our engineering team is working on a solution to cross-region load balance for App Engine Flexible and Standard. There is no ETA for the release but you can follow the Public Feature Request for updates.

[Hossein Amerkashi]

@Jordan,

The last response was on August 28, a month ago. Any updates on this? This is really urgent. THANKS 

[Jordan (Cloud Platform Support)]

There are currently no updates on the release date. Until there is a production solution released, creating your own proxy to load balance between regions is the only workaround. 

One workaround would be to setup a Cloud Load Balancer that accepted requests from your domain site. These requests would be routed to a Compute Engine instance in the requestor's region. The Compute Engine instance would simply act as a proxy and send the request to your separate App Engine project in their region. 

[Robert Dyas]

We need this too for GAE Standard Environment... basically a GAE Std in EU going to backend databases hosted on VMs in EU, and GAE Std in US going to backend databases hosted on VMs in US. We need both in the same project to share services accounts, datastore tables, storeage buckets, etc.

On Sunday, September 24, 2017 at 11:04:16 AM UTC-4, Hossein Amerkashi wrote:

[Hossein Amerkashi]

@Jordan,

For Cloud Load Balancer, this would mean deploying multiple instances of project in different regions?? 

[Kenworth (Google Cloud Platform)]

@Hossein

Yes correct. Since you can only have one App Engine project per Google Cloud project, the current workaround is to indeed copy your App Engine application to multiple Google Cloud projects in different regions. Then using Cloud Load Balancer with multiple Compute Engine instances in a single project, you would proxy to your different App Engine apps in their respective regions. 

[Hossein Amerkashi]

@Kenworth, Following this approach will possibly cause issue because all users are currently on central-usa region. If we deploy another instance for Europe, then their old data they have stored will not be accessible to them.

Could you or others, take a look at attached and see if it gives any info on performance issue?

Also, looking at our instance version, it says that version is deployed to F1. How can this be upgraded to +F4 OR is it auto-scaled dynamically??

THANKS,

-Hossein.

[Kenworth (Google Cloud Platform)]

The previous answer is a workaround available for GAE and it is not without issues. This route is usually taken by developers who wants to mimic the load balancing capability of Compute Engine into App Engine apps. The real solution is for your application to use Google Compute Engine and tap capabilities like global scaling, load-balanced computing, and the flexibility to configure your environment based on your needs.

When using automatic scaling, the F1, F2, F4 and F4_1G instance classes are available. You can change/upgrade your instance class by modifying the 'instance_class' element on your configuration file (app.yaml).