Azure Web App Service for Windows web App | DYNATRACE

[gsaharan]

Hi everyone,

I have a web application hosted on Azure App Service, and I’m planning to integrate Dynatrace for observability using the Azure App Service PaaS integration.

My objective is to get application-level to service-level visibility that can help us identify performance bottlenecks and improve the overall performance of the web application.

I’m familiar with the integration process, but I’m not clear on:

• What prerequisites or considerations need to be taken care of

• What level of application/service visibility we can expect

• Which key performance metrics and KPIs should be monitored

• What limitations or gaps we should be aware of with the App Service integration

I haven’t been able to find a document that covers these aspects end-to-end.

Could anyone please share relevant documentation or direct links covering Dynatrace observability for applications hosted on Azure App Service, including the expected metrics and monitoring capabilities?

Thanks!

[dylan_taelemans]

For service-, process-, and application-level monitoring, you will have access to largely the same metrics as you would when installing OneAgent in Full-Stack mode on a traditional server.

However, there are some limitations when it comes to host-level. For example:
- you won't have access to network metrics.
- Process instance snapshots are unavailable.
- for process you will probably only see your application process and no other host related processes.

If your main goal is application-to-service-level visibility, this approach should work fine.

As for prerequisites and considerations, make sure to disable Azure Application Insights. Otherwise, memory consumption can increase significantly as both dynatrace and Azure Application Insights will keep trying to inject, potentially causing the application to run out of memory and eventually crash.

[gsaharan]

Thank you for your inputs.

Also, is there any documentation available that includes snippets or examples of the metrics, which I can refer to and include in my documentation?

[gsaharan]

Also, I had one more question regarding the licensing. Could you please help me understand how it actually works?

[dylan_taelemans]

Licensing is based on the SKU/instance of your Azure App Service. It follows the same licensing model as Full-Stack Monitoring, with license consumption calculated based on the total memory of the SKU/instance type.

Regarding the available metrics, I'm not aware of any documentation that provides a complete list. However, looking at my own environment, I can at least already confirm the following host-level metrics:

• CPU usage -- 

  CPU idle

  CPU I/O wait

  CPU user

  CPU system

  CPU steal

  CPU other

• Memory usage --

  Memory total

  Memory reclaimable

  Memory used

• Page faults

• Container memory

All network and disk related metrics are available for me.

As mentioned earlier, for process-, service-, and application-level monitoring, you will have access to the same metrics as you would with a standard Full-Stack OneAgent deployment.