6.2 Dashboards and Widgets

Why Dashboards Matter on the Exam

Once the Agent collects metrics, tags, logs, and events, dashboards turn that telemetry into views teams can inspect, share, and discuss during incidents. The Fundamentals exam tests whether you understand what widgets do, how queries are scoped and grouped, and how reusable dashboard patterns like template variables reduce duplication. You do not need to memorize every widget type, but you must recognize the common ones and know how query controls change what a graph shows.

Dashboard Structure and Widget Types

A dashboard is a collection of widgets — visual building blocks that each query and display a specific data source. Widgets are not credentials, firewall rules, or Agent components; they are pure visualization layers on top of data Datadog already stores.

The timeseries widget is the workhorse. When an exam question asks which widget shows how a metric changes over time, the answer is timeseries — not Note, Host Map (which is spatial, not temporal), or Service Catalog.

Building Queries: From, Group By, and Aggregation

Every metric widget query has scope and aggregation layers:

Scope (the from box). In Metrics Explorer and widget editors, the from box filters which sources contribute data — for example env:prod, service:web, or host:db-01. Narrow scope first; querying everything produces noisy, misleading graphs. The from box is a filter control, not a notification editor, color picker, or site selector.

Group by. Grouping by a tag key such as host, service, or availability_zone splits one metric into separate series — one line per tag value. If you want one graph line per host for system.cpu.user, group by host. Without grouping, Datadog space-aggregates all matching sources into a single line.

Space aggregation controls how values from multiple sources combine at each timestamp: avg, max, min, or sum. Switching from average to max can expose worst-case behavior the average hides — a single saturated host buried in a fleet average.

Time aggregation (rollup) controls how points are combined within each time bucket (avg, max, sum over the interval). Space and time aggregation are independent dimensions; the exam may test one without mentioning the other.

Template Variables for Reusable Dashboards

Template variables (prefixed with $, such as $env or $service) make one dashboard work across environments. Instead of cloning separate prod and staging dashboards, you define $env as a dropdown and reference it in widget queries: avg:system.cpu.user{env:$env} by {host}. Changing the dropdown updates every widget that references the variable.

Template variables do not rename underlying metrics, disable rollups, or generate application keys. They are scope shortcuts — the exam frequently contrasts them with hard-coded tag filters.

Quick Graphs and Metrics Explorer

Quick Graphs is the fastest built-in way to visualize a metric without building a dashboard first. From almost anywhere in Datadog — a metric name, a monitor, a host panel — you can open Quick Graphs for ad-hoc exploration. If the question asks for the lightweight graphing workflow before committing to a permanent widget, Quick Graphs is the answer.

Metrics Explorer is the full query laboratory. It exposes from, group by, space aggregation, time aggregation, and function modifiers. One practical detail: the metric picker lists only metrics that reported in the last 24 hours. Quiet metrics still exist — type the full metric name manually to graph them. This 24-hour visibility rule is a common exam scenario.

Reading Timeseries: Partial Data Shading

On timeseries graphs, the newest point may appear shaded as partial data. This is not an error, expired API key, or invalid metric type — it means the current time bucket is still incomplete and Datadog is showing you data collected so far within that interval. Do not confuse partial shading with alert thresholds or archived data.

Dashboard Best Practices for the Exam

When the exam describes a team workflow, look for these patterns:

• One dashboard per service or team, scoped with template variables rather than duplicated boards.
• Timeseries for trends, Query Value or Top List for instant status, Table for multi-dimensional comparison.
• Consistent tag keys (env, service, team) so filters and group-by work across widgets.
• Quick Graphs first, dashboard widget second — explore before you commit.

Common Exam Traps

• Widgets visualize data; they do not replace the Agent, remove the need for tags, or create OS users.
• Template variables change query scope in the UI; they do not permanently rename metrics.
• Metrics Explorer's from box filters by tags; it does not edit monitor notification text.
• Changing a metric's type on the Metrics Summary page affects that metric globally across all dashboards and monitors — not just one open tab.