Spikes or steps in the telemetry of your app can now be easily diagnosed.
When you use Application Insights Analytics to explore app metrics over time, sudden changes, such as spikes or dips, are highlighted. With one click, Smart Diagnostics will find a pattern (a series of data with common values) that correlates with the change, and explain the reason behind it.
After you’ve created a time chart that includes an unusual change, click a highlighted data point. Smart Diagnostics finds a filter pattern that explains the data discontinuity – it identifies the pattern in which the discrepancy occurs, isolates it, and displays the result with and without the filter:
In this example, a new release of an app caused a sudden rise in the exception rate. Smart Diagnostics found a filter that characterized the change: The additional errors occur when users view a particular URL with a particular combination of browser version and client operating system. Smart Diagnostics displays the result with the filter false and true, so that you can easily see how strongly the change depends on the filter. In this case, with other combinations of URL, browser, or OS, the background trend continues as it was before, while all the additional errors are attributable to the problem values.
This information helps you focus your investigation on the root cause of the issue.
Want to know more?
Learn more about Smart Diagnostics in Analytics.
See a demonstration of Smart Diagnostics on sample data.
Watch Smart Diagnostics short video.
Smart Diagnostics is currently in preview, and we’re very keen to get your feedback – just click the smiley face icon at Smart Diagnostics results tab, or at top right of the Analytics window.
Quelle: Azure
Published by