Application Insights: More control of availability tests
Posted on 13 October 2015
Now in Application Insights, you have more control over availability web tests and you can reduce the number of a false-positive alerts. Retries enabled by default If a web test fails, this setting causes a retry within twenty seconds. The setting used to be disabled by default, but we discovered that in about 80% of cases, apparent failures went away on retry. So we've switched it to being enabled by default when you create a new test. Existing tests are unaffected. Custom timeouts for ping tests By default, our tests wait up to 120 seconds for your site to respond. If you want to be alerted about slow responses, you can now set a shorter timeout. More default test locations We choose five locations from which to send requests to your site. As previously, you can change the locations, choosing between one and 16 locations around the globe. Using several locations reduces the risk of network congestion being reported as a failure of your site. Alert sensitivity You can choose separately the number of locations that report failures and the period over which the failures should persist before an alert is sent. The default is 3 out of 5 locations fail for five minutes. Dependent resource files You can choose whether to download images, style files, scripts, and other dependent resources as part of a test. Test frequency Tests are run every five minutes from each location by default, but you can reduce the frequency to 10 or 15 minutes. For more information, see Availability web tests.