Exploring Data at Scale with New Capabilities in Application Insights and Metrics Explorer

Application Insights can capture, store and index large scale amounts of data about your application’s availability, performance and usage. But how can you gain insights from this data once it is stored in Application Insights? In this blog post, we will take you through some of the brand new features in Metrics Explorer for exploring large volumes of data, and extracting key insights from it with just a few clicks.

How to Open Metrics Explorer

In this blog post we will assume that you have an Application Insights resource configured in the Azure Preview Portal (if you are interested in setting one up, you can check out our videos or docs on the subject). Once you are set up, it is very easy to get to Metrics Explorer. Simply click on any chart in the overview blade that has data in it, for example:

This will open a curated Metrics Explorer view with charts and grids relevant to the chart you clicked on. From here you can begin your exploration.

Filtering Data in Metrics Explorer

Metrics Explorer now supports filtering data in the blade, via a new Filters button. Here we want to see requests coming from the United States, so we expand Country or Region and check United States. The Metrics Explorer blade then immediately updates to show the filtered data:

Once the filter is applied, a tag will appear in the main blade to indicate that the filter is active. You can also apply multiple filters by simply checking more boxes in the Filter blade. Now we want to see users from the United States who were using Firefox, so we expand Browsers and check Firefox:

You can also search through all of the Properties that are available to filter on by typing into the search box at the top of the Filter blade:

Customizing Reports in Metrics Explorer

All Metrics Explorer views, including the curated ones that are included by default in the product, are fully customizable. A number of new features in Metrics Explorer make this really easy to do.

Adding a New Chart or Grid

Starting with any report in Metrics Explorer, you can add a new chart to it by clicking Add Chart. This will add a new chart into the Paired Timeline (multiple charts with a common time axis) that you can customize:

Once you add the chart, you can select one or more metric – here we add Browser Page Load time in to compare it to the number of server requests and page views, and then add another chart for Server Response Time:

You can now view this data together with the rest of the data streams in your paired timeline. Hovering over values in one chart will also show you values from the other charts at the same point in time.

Now we are curious about the Failed Requests. How are these broken down by browser? To find out, we select that chart and turn on Segmentation, and segment by Browser. To make the chart taller, we adjust its height using the Chart Height setting. We see that virtually all of the failed requests are happening in Chrome:

To get a baseline, we turn on segmentation for Server Requests and see that most of the users are using Chrome there as well:

We are curious to see more about these, so we switch the chart to a grid by clicking on Grid in the Chart details blade with a single click:

We want to compare the total number of requests to the number of User Sessions and Server Response Time, so we select additional metrics in Chart Details which immediately appear in the grid as additional columns:

To see the individual values behind these, we can click on a row – here we click on Firefox to see the exact items behind that row:

Creating Your Own Report from Scratch

If you want to create a new report from scratch, you can click on the Metrics Explorer part in the overview blade to start out with a blank Metrics Explorer that you can make your own.

Visualizing Custom Metrics

It is really easy to visualize custom metrics in Metrics Explorer as well. For a more detailed look at custom metrics in Application Insights, see our previous post. Simply choose the metrics of interest from the Custom Metrics section of the Chart Details blade, and you are good to go. All of the features described in this post will work great for custom metrics as well as built-in ones.

Usability Improvements in Metrics Explorer

We have implemented a number of usability improvements in Metrics Explorer as well. Now, when you hover over a point in a Paired Timeline in Metrics Explorer, you can see the exact time of that location labelled on the hover bar. In addition, hovering on series or metrics that are unlabeled will show the metric labels in view on the right.

Metrics also now all have info icons that include descriptions of the metric, as well as details about how the data is aggregated and displayed. Metrics Explorer also now displays data from the prior week with a dotted line superimposed on charts, allowing you to compare the current time period to the same day from the past week as a baseline. This is also configurable.

Summary

In summary, we have seen how you can visualize complex data using key new capabilities in Metrics Explorer. Some of the key new capabilities in Metrics Explorer are:

- Filtering

- Improved Segmentation

- Customizable Grids

- Multi-metric Support in Grids

- Improved Grid Visualization

- Add Chart, and Chart Height Customization

- Changing Between Charts and Grids

- Usability Improvements

You can try them out today. Simply login to the Azure Preview Portal and create an Application Insights resource to get started.