15.2 Dashboards: Tiles, Variables, Segments & Continuous Monitoring
Key Takeaways
- A dashboard consists of tiles, and the four tile types are Explore, Query, Code, and Markdown.
- Explore tiles use a point-and-click interface over logs, metrics, and business events; Query tiles display data returned by a Grail query.
- Variables make one dashboard serve many contexts, and the four variable types are DQL, Code, List, and Free Text.
- Variables can filter the dashboard, supply values to code tiles, and act as placeholders in tile titles and markdown text.
- Custom segments can be enabled per tile from the tile edit panel, letting a single dashboard be re-scoped without duplicating it.
A dashboard is the platform's continuous monitoring surface. Section 15.1 established when to reach for it; this section covers how it is built, because the exam asks about tile types, variables, and segments by name.
Dynatrace's own positioning: "Dashboards are a cornerstone of the Dynatrace Platform, offering visual interfaces that provide real-time insights into application and infrastructure performance. By aggregating key metrics and data in a clear, customizable format, dashboards allow users to monitor, analyze, and troubleshoot issues."
The Four Tile Types
A dashboard consists of multiple tiles, and each is one of four kinds:
| Tile type | What it does | When to use it |
|---|---|---|
| Explore | Explore data such as logs, metrics, and business events with a point-and-click interface | Fast tile creation without writing DQL; ideal for less technical builders |
| Query | Displays data queried via Grail using DQL | Anything the point-and-click builder cannot express |
| Code | Displays data returned by code executed via Dynatrace functions | External sources, custom transformations, blended data |
| Markdown | Static content edited in markdown | Titles, section headers, runbook links, escalation contacts |
The Explore-versus-Query distinction is the one exam questions target. Explore is a guided builder; Query is raw DQL. A scenario describing a business analyst who must add a tile without learning DQL is answered with Explore.
The Markdown tile is under-appreciated in practice and worth remembering: putting the on-call escalation path and the runbook link directly on the dashboard is what makes it usable at 3 a.m.
Variables: One Dashboard, Many Contexts
Without variables, organizations duplicate a dashboard per team, per environment, and per region, and the copies drift apart within weeks. Variables solve this. Use variables to filter your dashboards, to act as variable values in code tiles, and as placeholders in tile titles and markdown tile text.
Those three uses are worth memorizing individually, because the placeholder capability is the one candidates forget: a tile titled Error rate — $namespace renames itself as the viewer changes the selection, so the dashboard never lies about what it is showing.
There are four variable types:
| Type | Value comes from | Typical use |
|---|---|---|
| DQL | The value is returned from a query | A live list of namespaces, services, or hosts that stays current automatically |
| Code | The value is returned from code | A list assembled from an external system, such as a CMDB |
| List | A comma-separated values list | A short, stable set such as prod,staging,dev |
| Free Text | Free text entered by the viewer | An order ID, trace ID, or customer identifier |
A DQL-backed variable is almost always the better engineering choice over a List variable for anything that changes: a hardcoded list of services silently goes stale the moment a new service is deployed, whereas the query-backed list picks it up on the next refresh. Exam scenarios that mention "new services keep being missed from the dashboard" are pointing at this.
Segments
Where variables parameterize a query, segments re-scope the data. You turn on Custom segments in the tile edit panel, select a segment from the Custom segments list, and add additional segments to the selected tile for filtering data.
The practical difference:
- A variable changes an input to the query — "show me namespace
checkoutinstead ofsearch." - A segment applies a saved, named scope — "show me the EMEA production estate" — consistently across tiles and reusable across dashboards.
Segments are how a platform team publishes one golden dashboard that each tenant reads through their own scope, which pairs naturally with the management zones covered in Chapter 12.
Designing a Dashboard That Gets Used
The scenarios on the exam usually reward the same design instincts:
- One dashboard, one audience, one question. An executive availability board and an SRE debugging board should not be the same artifact.
- Lead with the answer. Put the two or three tiles that state overall health at the top; put breakdowns below.
- Parameterize instead of duplicating. If you are about to clone a dashboard, add a variable.
- Add a markdown tile with the runbook. A dashboard that shows a red tile without saying what to do about it is only half-built.
- Only promote validated queries. From Section 15.1 — tiles should come from questions that proved useful in a real investigation.
Drilldowns
A dashboard is an entry point, not a terminus. Tiles support drilldowns into the underlying app — from an error-rate tile into the Logs app with the same filter and timeframe applied, or from a service tile into distributed tracing. The exam framing is that the dashboard tells you that something is wrong and the drilldown carries the timeframe and filter context into the app that tells you why, without the analyst having to rebuild the query by hand.
A platform team maintains eleven near-identical dashboards, one per product team, and they have drifted out of sync. What is the intended Dynatrace design pattern to consolidate them?
A dashboard uses a variable listing the services it monitors, but newly deployed services are consistently missing until someone edits the dashboard by hand. Which variable type was most likely used, and what should replace it?
A business analyst with no DQL experience must add a tile showing business event counts by region to an existing dashboard. Which tile type is designed for this?