Using the analytics dashboards
Analytics dashboards help you see and detect changes in your API ecosystem at a glance. The ability to see what has changed over time helps you identify problems and take corrective action quickly.
For a quick overview of Analytics Services, who uses them, and why, see Analytics Services overview.
This topic explains how some of the common features you'll find in all the dashboards. After reading this topic, you will understand:
- The general layout of all dashboards
- Common features all dashboards share and how to use them
- A few tricks and tips that will help you get the most out of the analytics dashboards
Check out this short video to learn about new API Analytics features and see how you can view trends, spot anomalies, see top performers, and compare different metrics in an API ecosystem.
Has there been a sudden spike or drop off in API traffic? Which app developers are most successful? What is the adoption rate of your API among developers? Which API methods are most popular? The Edge Analytics dashboards are designed specifically to answer questions like these.
In the background, Apigee Edge collects information as data passes through your APIs. The dashboards provide a powerful way to use this data immediately. If you see something of interest in a graph or chart, an anomaly or sudden change, you can then drill deeper to uncover as much detail as you require. If you notice that a particular developer is experiencing a lot of errors or a sudden drop in traffic, you can contact that developer proactively. Dashboards give you insight into your APIs that allows you to take action.
Most of the Apigee Analytics dashboards are organized and function much like the one below. If you understand how to use this dashboard, you'll be comfortable navigating the others.
- Data range and aggregation settings - Let you select the of dates for which to display data and the aggregation or "granularity" of data to display.
- Graph syles - All dashboards include one or more graphical charts, including line graphs, bar graphs, and pie charts.
When you select a smaller aggregation interval the larger the dataset the dashboard is working with. Performance tends to increase when you select a larger aggregation interval. For example, performance is greater if you select By Day rather than By Minute.
Figure: Typical dashboard organization
Dashboards have a set of common features, including time range and data aggregation settings, view togbgles, click and drag zooming on charts, mouse-over hover for more details on charts and other regions, and selectors for choosing the data to display in a chart. If you understand how to use one kind of dashboard, you'll be comfortable using the others.
- Toggle the view - Some dashboards let you toggle between distinct views or toggle between Detail and Summary views.
- Set the time range - You can select the time range over which the dashboard displays its data. Select Custom to use a calendar-style date picker to select the interval.
- Select the data aggregation interval - The aggregation interval ranges from minute, hour, day, week, and month.
Note: When you select a smaller aggregation interval the larger the dataset the dashboard is working with. Performance tends to increase when you select a larger aggregation interval. For example, performance is greater if you select By Day rather than By Minute.
- Zoom in - You can zoom in on chart data by clicking and dragging.
- Hover mouse over graphs - You can mouse over any point on a graph for more context about the data at that point.
- Hover mouse over labels - You can sometimes hover over a label for more details. Look for the "?" cursor as you mouse around.
- Export data to file - Some dashboards let you export CSV, PDF, or PNG data formats to a file. Look for the Export dropdown menu or the export icon:
The following figure highlights these feature areas:
Here's another dashboard that includes additional features that you'll see in some other dashboards:
- Refresh button - Updates the dashboard with the latest data.
- Export data to a file - Select the data format from the Export menu.
- Feature selections - Lets you show or hide certain features of the dashboard like investigating anomolies and showing moving averages.
- Dropdown menu for metric selection - Let's you pick the data dimension that you wish to plot.
- Plot selectors - Lets you select which set of data you wish to plot. In the example below, you can select which API proxy you wish to plot.
The following figure highlights each of these feature areas:
If you see a dramatic rise (a "spike") or a dramatic drop in the traffic, you can get further detail, by checking the Investigate Anomalies checkbox on the API Proxies page and clicking on a point in the chart that corresponds to the spike or drop.
Watch a short video to learn how the Anomaly investigation tool works.
In response, you'll be able to view the traffic pattern before, at, and after the spike or drop. You can display the Investigate Anomalies data by any of the dimensions available by default in Apigee Edge as well as by custom dimensions. This gives you enhanced insight into the cause of a spike or drop and enables you to correlate it to factors such as developer, developer app, resource, client IP address, or target URL.
In addition to the Investigate Anomalies checkbox, there are two checkboxes displayed for the APIs on the API Proxies page or for an API on its detail page.
|Show Moving Averages||Check this checkbox to view a moving average for the API. You can check this checkbox for multiple APIs to view a moving average that includes the set of these APIs. A moving average is a series of averages taken over sucessive subsets of a complete set of data. It's especially useful in viewing trends. The moving average is displayed as a band whose limits are +-20% of the calculated moving average data points.|
|Show Alerts||Select this checkbox (on the API Proxies page) to view the number of times that the moving average for the API exceeded the +-20% limit.|
When an analytics report illustrates an average, as well as minimum and maximum values, we display an accompanying dispersion box plot, as called out in the custom report below.
At a glance, a dispersion box plot enables you to see the central tendency and dispersion of your data. A dispersion box plot surfaces the five key numbers when it comes to illustrating averaged analytics data:
In this example, the area within the box indicates the most typical average target response times experienced by your traffic — 50% of your traffic, to be exact.
The line extending from the left side of the box indicates the average target response times experienced by 25% of your traffic.
The line extending from the right of the box indicates the average target response times experienced by the remaining 25% of your traffic.
The longer these lines, or “whiskers,” the more extreme your outlier values.
Help or comments?