Add Application Insight to a existing Azure Service Fabric cluster

Question

We have been working on Azure Service Fabric since last 4 months but now have decided to have a unified loging infrastructure for which we are using AppInsight, how can we a

Answer 1

I did it this way, using Visual Studio 2017:

1. Open Cloud Explorer from VS
2. Expand Virtual Machine Scale Sets
3. Locate the scale set containing your SF cluster.
4. Right-click Update diagnostics...
5. In the dialog, check Send diagnostics to Application Insights and add the Instrumentation key for you AI sink
6. Click OK and leave the dialogue
7. You might get an error (I did), but if you re-open the dialog you will find that the settings are still there.
8. To 'apply' the new settings to the cluster, open your azure portal, select All resources, and open the Virtual Machine Scale Set
9. At this point, hopefully selecting Restart will apply the changes to the cluster. I did a Deallocate followed by a Restart and a re-deploy of my SF-application, and this worked (I did not test if just doing a restart would work). However be very careful using this approach, make sure you know the trade-offs when de-allocation a SF-cluster

Answer 2

Unfortunately it is not just that simple right now.

The Application Insights support as shown at Build 2017 is still in preview. It consists of two nuget preview packages you have to install and you need to update your cluster to the latest version. Stay tuned to this repo which at the moment shows

Using Service Fabric with Application Insights

We are working on updating this repo to demonstrate how to use Service Fabric with Application Insights. The completed repo is not quite ready, please check back here within the next couple of days!

More information about the setup and nuget packages can be found here:

This repository produces the following two nuget packages:

Microsoft.ApplicationInsights.ServiceFabric.Native - For use with Service Fabric's native reliable services. Microsoft.ApplicationInsights.ServiceFabric - For use with Guest Executable and Guest container services (lift and shift scenarios).

Until that time you can take a look at this repo for full (manual) Application Insights integration.