Get alerts of performance or availability issues

Visual Studio 2013

Application Insights is in preview.

If your web application is performing badly—or is not running at all—you’d like to know about it immediately. Microsoft Application Insights monitors your web services at regular intervals, and can send you an email if performance drops below an acceptable level. Also, the alerts page shows you at a glance any issues that you should be concerned about.

You can set up two kinds of alert:

  • Availability alerts tell you if your website responds slowly or not at all

  • Metric alerts tell you if a performance counter (such as CPU usage or uncaught exceptions) crosses a given threshold.

When you create or edit an availability test, the dialog has a field where you enter the alert criteria. If you’ve chosen to run the availability test from several locations around the world, you can specify how many have to fail before you get an alert.

Set locations, content test, and alert rule

When you’ve set up performance monitoring for your web app, you’ll find metrics and events showing up on the Diagnostics pages.

On the Metrics page, select your application. Edit the list of metrics to show the one you want. Then select it and configure an alert.

Under Diagnostics, Metrics, select a metric

All the alerts for an application are sent to the same email address.

In the dialog to set up an alert, follow the link to configure notifications.

Make sure you’re working with the right application, and set the notifications address.

Check the option to send emails.

You only need to set notifications once for each application.

In the Overview/Alerts page, there’s a row for every alerting rule in your account. You can filter to one application component if you want.

Each row shows a history of the recent evaluations of a rule. The Since column shows how long the current state has persisted.

Choose Overview, Alerts, then select a row.

Some things you can do:

  • Change the date range—up to 2 weeks.

  • Widen the History column to see the details more easily.

  • Filter by application name, metric or test name. Enter a substring of the name and the results will be filtered as you type.

  • Select a rule to see its recent evaluations:

    Choose Overview, Alerts, then select a row.

  • Drag across part of the smaller graph to zoom in on any section. The heights of the markers are not significant.

  • Click any marker to see details of that instance.

  • Click the metric link to see the source metric:

    Follow the link to the metric.

On the Alerts page, choose Manage rules.

Overview>Alerts; Manage>Manage.

Choose your app and then select a rule to edit or delete.

Pick your app and then select a rule to edit.

Alerting rules are stored in a configuration file that you can download and upload, so that you can do batch updates. Your synthetic monitor availability tests are defined in the same file.

  1. Open the settings page.

    Choose Administer Account at top right
  2. In the Configuration tab, select your application, select the latest configuration, and then download the configuration file.

    Choose latest change, then Download

    Notice that every change is listed, including alerts and availability tests that you configured manually.

    The complete configuration is downloaded—not just the changes.

  3. Edit the configuration file.

  4. Refresh the list to make sure that no changes have been made since your download.

  5. Upload your updated file.

Q: I don’t have any items in my configuration history. How can I get a file to edit?

A: Configure an alert manually. That will create a configuration file that you can download.

It’s good practice to download an existing configuration, edit it, and then upload the new version.

  • <SyntheticMonitors> is the list of availability tests.

  • <Observers> is the list of alerting rules of all types.

You can add multiple rules in each configuration file.

These metrics are measured by the Microsoft Monitoring Agent.

Metric name

What it measures


Application performance events per second


Application exception events per second


Average request time in milliseconds


Monitored requests per second


Processor occupancy percentage


Disk IO bytes per second


Network bytes per second


Memory bytes





These metrics are measured by Analytics synthetic monitors, which test the responsiveness of your web service from multiple locations around the world.

Metric name

What it measures


Average availability raw


Average availability

Was this page helpful?
(1500 characters remaining)
Thank you for your feedback
© 2014 Microsoft