5 tips to build a better dashboard

Collecting metrics from distributed systems spanning private datacenters, public clouds, and a variety of third-party services has led to a data deluge for IT operations teams. When an alert is triggered, you don’t want to waste precious minutes sifting through the data trying to figure out what is happening. You need an effective way to visualize all of this data. That’s where dashboards come in.

Dashboards provide a quick way to see how systems and applications are performing. Dashboards reveal critical information about production data, system performance, business metrics, and user satisfaction. A dashboard is only useful when it can help users quickly see what is happening. Information and visualizations should be presented in a way to reduce cognitive load and toil for those viewing the data.

The challenge with dashboards is we often end up summarizing information, or displaying giant text tables, which leads to a loss of detail. This, in turn, results in greater cognitive load. Crunching massive amounts of data can take time, which results in shortcuts such as aggregating data for faster display. Or if you don’t want to aggregate data you create multiple dashboard views. But then you get lost looking for the right dashboard or switching between multiple dashboards.

Questions arise when examining visualizations on a dashboard. We need to easily follow the data to answer these questions in as few clicks as possible. Dashboards have long been a part of the Catchpoint solution. Whether it is the default dashboard, a customized dashboard, or Smartboard our goal is to help you separate the signals from the noise.

This week we unveiled our new default and node dashboard to address these challenge and help you find answers faster. When designing the dashboards, we followed these five rules to build an ideal dashboard that displays a summary of key alerts, tests, and global measurements of your end user experience.

Let’s dig into each of the rules and how it applies to the dashboard

1. Focus on the end user, not the data.

We thought about what our users are trying to accomplish when they log into the Catchpoint portal. Most users access the portal to check the health and status of tests that are running. If something is wrong, they quickly want to be able to follow the data to an answer. That’s why on the default dashboard we display an overview of errors and alerts received in the last hour across all tests.

2. Make it easy to drill down into data.

It’s not enough to know there is an error or a problem in a location. You need to be able to quickly drill down into what is going on. Mouseovers provide pop-ups with additional information and further drill downs to the node dashboard or detailed waterfalls. Getting you to the details you need with a single click.

3. Utilize insightful visualization

Visuals are easier for the brain to decode than a wall of text. We’ve added a node map showing the aggregate health of each test at the nodes you are using.

From the node map, you can drill down into a detailed dashboard for each node with additional visualizations.

The error summary displays the aggregate number of errors across all tests for the last hour. Providing visuals into whether things are improving, getting worse, or staying the same.

4. Group data logically and use space wisely.

Since we determined most users access the portal to investigate an error, the Error Summary, and Recent Alert sections are right at the top. Below that are more granular visualizations on the state of each node and data from Real User Monitoring.

At Catchpoint we have frequently said that Synthetic Monitoring and Real User Monitoring are better together, but our dashboards did not support this statement. There was one dashboard for Synthetic tests and a separate dashboard for RUM tests. Our latest release remedies that.

5. Less is more

It is easy to get carried away when creating a dashboard. The desire to include widgets for all the things can be overwhelming. But that will increase the cognitive load on users. Chances are they are already in a heightened sense of awareness as they are searching for answers to a problem. There is no need to add to the cognitive load and toil with endless visualizations. We’ve kept the data on the default dashboard to a minimum (don’t worry the ability to customize and add additional widgets is coming). We limited the number of visualizations and widgets to make the data easier to process.

If you haven’t logged into the portal recently, what are you waiting for? Go check out the new dashboard.