The key to getting started with dashboards is knowing what kind of questions you ask yourself regularly. What are common issues your customers face? When a problem occurs, what questions help you find a solution?
Creating a good dashboard is about bringing the answers to these questions to the surface. Also, it is important not to cram all of those thoughts into the same dashboard. Creating separate dashboards to pinpoint different issues can help you quickly find your answers.
This guide gets you started on a path to creating dashboards. These basic dashboards enable team discussion and speed up issue resolution.
If you haven’t already, create a Datadog account. Install the Agent on a host, and an integration for something running on that host.
Determine the purpose of the dashboard you’re creating. A dashboard can help you and your teammates focus on the right work. A team dashboard reminds you what’s high priority, what needs attention, and what you’re succeeding at. Make a team dashboard (or multiple) with the information that people most frequently need they have to dig for. SLO and SLI details make for an excellent team dashboard.
A dashboard connected to real-time data is a powerful tool for guiding conversations with managers and executives. A good executive dashboard can show that you are working on the most important things, how much a service is costing you, or whether you’re progressing toward goals, meeting your SLOs, and scaling effectively. Executive dashboards are most effective when they answer these questions at the highest level and are interconnected to drill down into the answer.
Dashboards can also help you track down persistent problems and fix them. Troubleshooting dashboards often start as a scratch pad of things you know, and gradually build as you discover more. For example, start with a graph or widget from another dashboard or view that shows a problem. You can then drill down from there to find your solution.
Datadog provides many out-of-the-box dashboards for features and integrations. For the infrastructure you monitor, check out the out-of-the-box dashboards that are provided with Datadog:
Redis, or a feature you use, such as
A common way to start a dashboard is by encountering a similar dashboard already in use, and adjusting it to suit your needs. If you find a dashboard that answers many of the questions you want your dashboard to answer:
Command + C(
Ctrl + Con Windows). Paste it into your dashboard by opening the dashboard and typing
Command + V(
Ctrl + Von Windows).
Through integrations, Datadog collects metrics from your infrastructure and applications. The collected metrics are documented in the integration’s README files. If you encounter a metric in the Metrics Explorer or as you’re creating a dashboard, and you want to know what the metric is, look it up in the Integrations docs.
For example, suppose you are looking at a time graph of the metric
aws.s3.first_byte_latency. Go to the Data collected section of the AWS S3 integration README to see its description:
The average per-request time from the complete request being received by a bucket to when the response starts to be returned. Shown as millisecond.
After you’ve selected a few metrics to add to your dashboard, experiment with various widget types, queries, functions, and aggregation approaches, to display the data in ways that best answer the questions you have.
By specifying Template variables, you can make one dashboard answer questions for a selection of scenarios. For example, you can create a time graph that shows latency metrics for whichever data center geography the user selects from the dashboard’s variables drop-down, or for all of them together. For more information, see Template Variables.
You can make graphs easier to read by adjusting Y-axis ranges, colors, or legends, or by adding markers and event overlays. See the Dashboards documentation for all the ways you can customize and refine timeseries and other widgets.
For more details and examples of these techniques, sign up for the online learning course Building Better Dashboards.
Timeseries graphs of metrics are useful, but dashboards can contain many types of widgets to communicate important information. Try:
See Widgets for more information and examples of setting up these graphs.
Move graphs around so they create a flow for the work you do or conversations you have around the dashboard. Drag and drop widgets to place them. On screenboards, use Free Text widgets to organize sections under headings. On timeboards, add a Group widget that can contain multiple widgets, and can collapse out of the way when you’re viewing the dashboard.
There are two ways to create links from a dashboard to any target URL:
Click Generate Public URL in a dashboard’s Settings (gear) menu to create a URL you can share with big screens or people who don’t necessarily have a Datadog account.
Integrate with your team communications by using the Slack integration to import dashboards and other Datadog features, such as monitors and incidents, into a Slack channel.
Every dashboard has a JSON representation that you can copy or export from the Settings menu. Each widget on the dashboard also has a JSON definition, which you can see and edit by opening the widget editor (pencil icon) and clicking the JSON tab under Graph your data.
Because all widgets and dashboards are represented as JSON, you can programmatically generate them by using the Dashboards API, which is useful if you want to generate a Dashboard every time your team starts a new project or encounters an incident, or formalizes an SLO, for example.
Additional helpful documentation, links, and articles: