search

MicrosoftDocs / azure-docs

Open source documentation of Microsoft Azure
https://docs.microsoft.com/azure
Creative Commons Attribution 4.0 International
10.25k stars 21.42k forks source link

Conflicting information with other pages, clarity on recommended practice? #25495

Closed ththiem closed 5 years ago

ththiem commented 5 years ago

It is a bit unclear what is necessary, and recommended, for diagnostics and monitoring solutions for service fabric. The information here seems at odds with other pages.

For example, this page states that you only need to create the app insights resource in azure and then add the fabric native nuget package to your services and it gives full comprehensive telemetry out of the box, organized and stamped by types and other properties.

Compare this to what's stated on the page under How-to Guides > Monitor and diagnose:

https://docs.microsoft.com/en-us/azure/service-fabric/service-fabric-diagnostics-event-analysis-appinsights

Where it says nothing about adding it to the code or nuget packages, but requires you to set up diagnostics on the fabric cluster itself through azure portal when it is created, and/or through the resource manager template by use of Sinks, and also states that all logs will only show up as traces and not as any other type. It then also goes on to recommend that EventFlow be used as the preferred solution. That article is also a fair bit older than this (current) one

Also of note, that the other article was written before .NET Core 2.2 was released, which claims to have brought feature parity with .NET Framework for Application Insights SDK, which is not clear in this article if that is a requirement to leverage Application Insights in the manner as described in this document.

Please advise. Cheers.

Document Details

⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.

dkkapur commented 5 years ago

@srrengar

srrengar commented 5 years ago

hi @ththiem thanks for the feedback. You are correct, that article is a bit confusing, The recommended practice is using the App Insights SDK to instrument your application, with this you get application level diagnostics out of the box. However, you can also see platform events from Service Fabric (NodeUp, Down, etc.) in App Insights too and that is done through the portal configuration. That section of the article should be under "monitor clusters" which I will change in the docs. Thanks for pointing this out and hope this helps!

ththiem commented 5 years ago

Thanks for the quick reply. Can you also confirm, for .NET Core, if 2.2 is required for App Insights SDK features in Service Fabric, as mentioned here?

https://docs.microsoft.com/en-us/azure/azure-monitor/app/asp-net-core#application-insights-sdk-comparison

srrengar commented 5 years ago

@brahmnes could you comment on the .NET Core version required for the App Insights SDK for Service Fabric? I didn't see it in the repo https://github.com/Microsoft/ApplicationInsights-servicefabric

brahmnes commented 5 years ago

The Application Insights for Service Fabric nugets doesn't require anything specific in .NET core 2.0. It should work with anything from .NET core 1.1 and above. However, I am not sure about Application Insights SDK itself. I guess it depends on which version of the Application Insights ASP.NET core nuget and which version of Service Fabric SDK you are using.

You can check on the dependencies for the nuget. For example: https://www.nuget.org/packages/Microsoft.ApplicationInsights.AspNetCore/2.6.1

ththiem commented 5 years ago

Sounds good, thanks for the advice. Closing this.