question

BoopathyElango-7282 avatar image
0 Votes"
BoopathyElango-7282 asked ·

Azure app service capacity

Hi ,

We have a third party application (WAR distribution) has to be deployed into Azure App service. Predominantly we are planning to use Tomcat 8.5 & Java 1.8 as platforms of our choice. Currently the application is running in on-primes data center with capacity of 4 CPU and 16 GB RAM with multiple instances of app running behind load balancer (LB).

Therefore we would like to know how much maximum and minimum threshold size of CPU and RAM the Azure App service supports today to run an application ? What is the memory footprint supported ? The application has setup configured for HTTP Session to store user identity. Is Azure app service supports sticky sessions and whether the user session will be distributed across multiple scaled instances ?

Moreover if we like to deploy the same application using Azure app service as containers (Docker container image), can we run application with the similar capacity size in containers ? What about sticky sessions and scaled instances in containers , is it supported ?

Thanks,
Elango Boopathy.

azure-webapps
10 |1000 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

1 Answer

JasonFreeberg-3971 avatar image
0 Votes"
JasonFreeberg-3971 answered ·

Hi,

It looks like a P3v2 will be the hardware tier closest to what you're currently using. App Service optionally supports sticky sessions (more info). If possible, you should store the user information in a remote Redis cache. This distributes the session information to all instances. The problem with sticky sessions is that it can cause an uneven distribution of traffic to your instances. Everything I said here applies to deploying containers and/or code.


· 2 ·
10 |1000 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

@JasonFreeberg-3971 ,

Thanks for clarifying the point. If in case the instance scales up due to spike in traffic then the one more instance with hardware tier P3v2 will be provisioned as per a scaling policy.

Could you clarify how to optimize cost using hardware ? So should I examine how much max CPU and RAM is used in each PROD server instance during the peak load ?
Whats the major difference between P3v2 and PC3 instance. I believe latter provides more RAM with less hourly rate compare to former.

Thanks

1 Vote 1 ·
ryanchill avatar image ryanchill BoopathyElango-7282 ·

See https://docs.microsoft.com/en-us/azure/virtual-machines/sizes-general#dv3-series-sup1sup for more information. But from a hardware standpoint, P3v2 uses Dv2 and PC3 uses Dv3. Although PC3 is tailored towards container images with faster SSDs and a higher Memory to CPU core ratio.


1 Vote 1 ·