The Rise of Drop Failure Rate and Unclassified Streams: Explained

This post has been republished via RSS; it originally appeared at: Microsoft Teams Blog articles.

As customers start transitioning from Skype for Business to Microsoft Teams for calling and meetings, a lot of attentions have been given to understanding how Teams behave in their environment where Skype for Business has been operating previously.

 

Call Quality Dashboard (CQD) is the tool of choice that allows us to gain insights into the reliability and quality of calls made using Teams over a period of time, and you can also use Call Analytics for a point-in-time, call reliability and quality troubleshooting efforts. We have also published the Quality of Experience Review (QER) guidance that complements CQD by providing a comprehensive guide to leverage it to achieve and maintain high quality experiences for your end users.


Looking at the several customers’ data from the past six months, we have noticed a trend of increased audio drop failure rate in the order of extreme magnitude and unclassified audio stream counts affecting several customers that we’ve been working with, especially from the middle of November onwards.

 

CQD-1.pngCustomer 1 Summary Report

 

CQD-2.pngCustomer 2 Summary Report

 

Based on our investigation, the dropped calls and ‘unclassified’ calls are inactive audio sessions in a screen sharing only, two-person ad-hoc group call. There is no real user impact, but there is an artefact due to how the feature was implemented, which was the extra dropped calls and ‘unclassified’ calls.


We realize accurate telemetry is critical for our customers’ operations, therefore, we will deploy a change to exclude these sessions from the Call Quality Dashboard by February 1st, 2019, and in the coming weeks the fix will be implemented in Call Analytics as well.

 

(Kudos to for the collaboration to co-author this blog post)

Leave a Reply

Your email address will not be published. Required fields are marked *

*

This site uses Akismet to reduce spam. Learn how your comment data is processed.