How can we improve Application Insights?

Add support for Intranet application testing (not just via Internet). E.g., via NTLM

App Insights works great for externally-facing applications. Would like the same rich experience for Internal/Intranet applications.

103 votes
Sign in
(thinking…)
Sign in with: Microsoft
Signed in as (Sign out)

We’ll send you updates on this idea

Lionel Godolphin shared this idea  ·   ·  Flag idea as inappropriate…  ·  Admin →

With the release of Azure Functions 2.0, the previously posted template no longer works.

However, we recognize the need for this request, and are actively scoping and prioritizing the work needed to make it happen.

I am very interested in speaking with anyone who needs this feature to learn more about your scenarios, particularly any information around cost issues. If you would like to help me scope this feature, please reach out: mogrobin@microsoft.com.

Morgan Grobin
4/26/2019

16 comments

Sign in
(thinking…)
Sign in with: Microsoft
Signed in as (Sign out)
Submitting...
  • Peter Hewitt commented  ·   ·  Flag as inappropriate

    Given our company is not going to be exposing our applications externally, but will have a number of services/micro services that using application insights along with health checks would be helpful. However all the performance and testing aspects of Application insights requires externally facing sites or server less functions.

    As we start to implement Health checks in our asp core services we would like to make sure our availability can be tested and alerted on. All this without have to open up firewalls.

  • AdminApplication Insights Team (Product Manager, Microsoft Azure) commented  ·   ·  Flag as inappropriate

    The team has had to shift priorities, and we expect to complete the remaining UI pieces, and finalize the billing model in the April timeframe. Meanwhile, the TrackAvailability SDK is now available in the .NET stable SDK (https://www.nuget.org/packages/Microsoft.ApplicationInsights/).

    Please note that when the feature is released as public preview, we may limit the number of tests, and the rate at which you can report results from custom tests. Also, while this will be free for you to try in public preview, there may be a billing implication for continued usage when this feature is generally available.

    Please note that pending the UI changes, data sent via the TrackAvailability SDK will *ONLY* be accessible via the following:
    1. In the Metrics Explorer, the 2 Availability metrics - Availability and Test duration, will be aggregates across all your tests, including custom (reported via SDK) and the AI executed ping/multi-step tests. You may segment these metrics by test/location etc.
    2. In the Search Explorer - you can view raw results from all your tests
    3. Using Analytics queries: All results will be available for querying
    4. Alerts: The aggregate availability, and test duration metric will be available in the metrics dropdown, and include the results from custom tests.
    5. Continuous exports

    The 2 key areas, custom tests will NOT be available from are the following:
    1. The Availability exploration and diagnostics experiences.
    2. Alerting on availability of specific test locations

    Thanks,
    Dash

  • Vincent Boisclair commented  ·   ·  Flag as inappropriate

    This cruelly missing feature will be greatly welcome as it will fill the hole left for the apps locked in VNets. Definitely looking forward to this and news about the upcoming availability.

    The SDK approach is an ok workaround but enabling access to the VNet (similar to what was done for the App Service) would be even better.

  • James commented  ·   ·  Flag as inappropriate

    Looking for NTLM support because we are looking to replace Site 24x7 with Application Insights.

  • Dave Walters commented  ·   ·  Flag as inappropriate

    This is going to be a big value to mixed enterprise IT who have both externally-facing services, SaaS dependencies and internal dependencies and sites. Can a modification of telemetry endpoints accommodate this? Also, what about binding to authentication data (Ex, you have an Active Directory SSO method, bind your test to encrypted data and validate Intranet calls?)

  • Kaushik Patel commented  ·   ·  Flag as inappropriate

    As many applications have hybrid Infra (OnPrem + Cloud Services) it is very difficult to manage application availability at single place. It will be nice to have this feature.

  • Antonio Blanco commented  ·   ·  Flag as inappropriate

    Soubhagya,

    The SDK idea sounds like something that definitely works for my use case. In particular our company has a lot of internal sites like http://mysite, and giving us a way to report results to AI directly is 90% of the value add, especially for alerting on a failed availability test.

    For context, w/o a way to report to AI, we end up having to build out a reporting infrastructure, deal w/ data storage and reporting, and do regular SQL maintenance.

    In my case, I wouldn't mind installing something on a server behind my firewall.

    My prioritized list of features:
    1) The bare minimum would be to get the SDK. I can then add them into my web test.
    2) Next would be an AI WebTest Plugin, where I simply provide an iKey and maybe some configurations in Visual Studio so that I can add them to my webtests.
    3) A really cool experience would be something similar to status monitor, I simply download it onto a machine and configure via iKey
    4) If I'm running webtests on my machines, perhaps loosen the webtest restrictions to let me add my own plugins, or run coded webtests. We have a few scenarios where plain scripted webtests don't quite hit the software the right way.

  • Anonymous commented  ·   ·  Flag as inappropriate

    Hi,

    Is it possible now to use Application insights for tracking usage of application which does not have round the clock connectivity?

    I need tot rack the usage of a WPF application which will have internet connectivity only intermittently. Is it possible to track the usage when machine is not connected by allowing some bulk upload (from offline usage) to Azure when there is connectivity?
    Please mention if there is any limitation in case this is possible.

  • Dave Lubash commented  ·   ·  Flag as inappropriate

    The scenario Chuck Lazarr has below would work today if the devices are connected to the Internet. Supporting all of these scenarios, for Intranet Applications is a common ask, that we have this on our backlog, however, we are working to deliver great experiences for Internet first. As we get closer to delivering the Interanet experience, we will post an update.

    Thank you for your interest in Application Insights.

  • Anna commented  ·   ·  Flag as inappropriate

    Hello,

    Thank you for your suggestion.

    While we realize an availability monitoring solution for non-internet facing applications is important and something we both want and plan to do.

    However, this is will not make it into our Summer update. Hoping for the Fall or Winter App Insights updates.

    The reason for this prioritization is that we want to improve our existing monitoring capabilities for cloud-based applications first.

    Thanks,
    Anna

  • Chuck Lazarr commented  ·   ·  Flag as inappropriate

    I would like to see Usage (of a client application), at a minimum, for my scenario. We deploy employee-facing applications as well as Visual Studio extensions (to developers). Installation and feature usage statistics for these internal, non-web .NET-based applications would be quite valuable in our organization.

Feedback and Knowledge Base