none
App Insights running on windows App Service Plan intermittently stops writing telemetry to Log Analytics workspace. RRS feed

  • Question

  • Hello Forum Support Team,

    We have a couple of busy Web APIs that intermittently stop writing telemetry to the Log Analytics workspace.  We use App Insights Metrics & Log Search Queries for our Monitoring and this throws off our Metrics and breaks our Alert Rules.

    - This has resulted in over 30 minute gaps in our performance and monitoring data.   

    - It seems to occur during heavy workloads.

    - Mitigation Attempted: Restarting the specific Web App does mitigate issue

    - Workaround: Stopping Web App, waiting 3 seconds, starting the Web App, wait 5 minutes -> Telemetry now writing back to Log Analytics workspace.

    The workaround above will not work for us as we use App Insights Metrics & Log Search Queries.

    Please let us know if there is any additional logs or information I could provide to help get this issue resolved.

    Thanks,

    Gary


    Gary Gallanes



    Wednesday, August 14, 2019 9:42 PM

All replies

  • Hi Gary,

    There are a couple bits of information that would be helpful to start with:

    How long ago did the issues start?

    Is this occurring across all instances, or are some working and others not?

    What is the type of the instance? aka Linux/Windows, using Docker, etc.

    Is there any relevant information in the Application or Server logs?

    Thank you!

    Thursday, August 15, 2019 4:14 PM
    Moderator
  • Hi Samara,
    Thanks for the reply.

    Additional Info: 
    - We noticed the 'perfromanceCounters' will stop logging but other logging such as 'requests' sometimes continue to log. Other times all logging stops. 

    - I was mistaken, stopping and starting the Web Api does NOT fix App Insights logging. 

    - App Insights eventually starts logging again on it's own. Don't know why!

    Answers to your questions below:

    How long ago did the issues start? 
    Not certain. We deployed App Insights metric alerts for the first time 4 days ago. 
    We discovered the issue validating the alerts in Log Analytics.  

    What is the type of the instance? aka Linux/Windows, using Docker, etc.
    Issue is with App Insights in PaaS:
    ASE > Isolated App Service Plan > App Services on Windows Server > NodeJS API Instance

    Is there any relevant information in the Application or Server logs?
    The App Service logs show the Web API is up and responding to requests during the time App Insights 
    was not collecting or sending Telemetry. 

    We just enabled verbose diagnostic Application & Web Server Logging. 
    We'll respond in a couple of hours with any clues we find in the logs.

     Thanks,
    Gary

       

    Gary Gallanes




    Thursday, August 15, 2019 9:53 PM
  • Gary,

    Thank you for the additional information. Intermittent issues like this are always hard to troubleshoot, and nothing so far is jumping out at me as the cause.

    Could you send an email to AzCommunity[at]microsoft.com with your subscription Id and a link to this post? That will allow me to access some additional resources for you to assist with troubleshooting.

    Friday, August 16, 2019 3:23 PM
    Moderator
  • Hello Samara,

    We sent email to AzCommunity[at]microsoft.com with the requested information.

    Please let us know if there is anything we can do or additional information we could provide to help get this issue resolved ASAP.

     Thanks,

    Gary


    Gary Gallanes

    Monday, August 19, 2019 5:02 PM
  • Hi Samara,

    We've created our Service Request with Microsoft support and will post the resolution and mark this Question 'Answered' once completed.

     Thanks,


    Gary Gallanes

    Tuesday, August 20, 2019 8:32 PM