dt-app-dashboards（dt 应用程序仪表板）

Dynatrace Dashboard Skill

Overview

Dynatrace dashboards are JSON documents stored in the Document Store. Each dashboard contains:

• Tiles: Visual components displaying markdown content or data visualizations
• Layouts: Grid-based positioning (20-unit width) defining tile placement
• Variables: Dynamic parameters ($VariableName) for query filtering
• Configuration: Version metadata and dashboard-level settings

When to use this skill:

• Creating new dashboards with skill-based query generation
• Modifying existing dashboards (queries, tiles, layouts)
• Querying dashboard JSON to extract data or analyze structure
• Analyzing dashboard purpose, metrics coverage, and health

Four main workflows:

1. Creating - Build dashboards with skill-based queries
2. Modifying - Update tiles, queries, layouts, variables
3. Querying - Extract data from dashboard JSON
4. Analyzing - Understand structure, purpose, gaps, and health

Dashboard Document Structure

Dashboards in the Dynatrace Document Store include both metadata and content:

{
  "id": "dashboard-abc123",
  "name": "My Dashboard",
  "type": "dashboard",
  "owner": "user-uuid",
  "version": 60,
  "modificationInfo": {...},
  "content": {
    "version": 21,
    "variables": [],
    "tiles": {...},
    "layouts": {...}
  }
}

Metadata (top-level):

• .id - Document ID
• .name - Dashboard name
• .owner - Owner UUID
• .version - Document version (change tracking)
• .modificationInfo - Creation/modification timestamps

Dashboard content (.content):

• .content.version - Dashboard schema version (current: 21)
• .content.tiles - Tile definitions
• .content.layouts - Tile positioning
• .content.variables - Dashboard variables

All jq examples in this skill use the .content.* paths.

When to Load References

This skill uses progressive disclosure - load only what you need:

• Start here: SKILL.md provides core concepts and quick-start examples
• Load references on-demand: Each reference file covers a specific deep-dive topic
• Context efficiency: Progressive loading enables task completion without external documentation

Loading strategy:

1. Try answering with just SKILL.md first
2. If you need detailed specifications or advanced patterns, load the relevant reference file
3. The "References" section below maps each file to its use case

💡 Tip: Reference files are linked throughout this document with → arrows pointing to when you should load them.

Working with Dashboards

For detailed workflows and mandatory requirements:

• Creating & Updating dashboards → Load references/create-update.md for complete workflow, skill-based query generation, validation, and modification patterns
• Analyzing dashboards → Load references/analyzing.md for structure analysis, health assessment, and JSON extraction

⚠️ MANDATORY for creation/modification:

Follow this exact order (do not reorder):

1. Define purpose and load required skills, references and assets
2. Explore available data fields/metrics
3. Plan dashboard structure: logic, variables, tiles and layout
4. Design and validate all variable/tile DQL with dtctl query "<DQL>" --plain
5. Construct/update dashboard JSON
6. Validate the dashboard JSON structure and queries
7. Deploy the dashboard via the Dynatrace API

Full requirements and examples: references/create-update.md.

Dashboard Structure

Required Structure

{
  "name": "My Dashboard",
  "type": "dashboard",
  "content": {
    "version": 21,
    "tiles": {},
    "layouts": {}
  }
}

Optional properties inside content:

• variables - Array of dashboard variables (filters/parameters)
• settings - Dashboard-level settings (grid layout, default timeframe)
• refreshRate - Dashboard refresh rate in milliseconds (e.g., 60000)
• gridColumnsCount - Number of grid columns (default: 20)
• annotations - Array of dashboard annotations

Structure concept: Variables define reusable parameters, tiles contain content/visualizations, layouts control positioning. Each tile ID in tiles must have a corresponding entry in layouts.

Tiles Overview

📊 For detailed tile specifications, visualization settings, and query configuration, load references/tiles.md

Markdown tiles: {"type": "markdown", "content": "# Title"} Data tiles: {"type": "data", "title": "...", "query": "...", "visualization": "..."}

Visualizations:

• Time-series (MUST have time dimension via timeseries/makeTimeseries): lineChart, areaChart, barChart, bandChart
• Categorical (no time dimension, summarize... by:{field}): categoricalBarChart, pieChart, donutChart
• Single value / gauge (single numeric record): singleValue, meterBar, gauge
• Tabular / raw (any data shape): table, raw, recordList
• Distribution / status: histogram, honeycomb
• Geographic maps: choroplethMap, dotMap, connectionMap, bubbleMap
• Matrix / correlation: heatmap, scatterplot

→ See references/tiles.md for specifications

Layouts Overview

📐 For complex layout patterns, grid system details, and positioning examples, load references/layouts.md

Grid: 20 units wide. Common widths: Full (20), Half (10), Third (6-7), Quarter (5) Properties: x (0-19), y (0+), w (1-20), h (1-20)

Example: {"1": {"x": 0, "y": 0, "w": 20, "h": 1}, "2": {"x": 0, "y": 1, "w": 10, "h": 8}}

→ See references/layouts.md for patterns

Variables Overview

🔧 For detailed variable configurations, replacement strategies, multi-select, and limitations, load references/variables.md

Definition: {"version": 2, "key": "ServiceFilter", "type": "query", "visible": true, "editable": true, "input": "smartscapeNodes SERVICE | fields name", "multiple": false, "defaultValue": "*"} Usage (single-select): fetch logs | filter service.name == $ServiceFilter Usage (multi-select): fetch logs | filter in(service.name, array($ServiceFilter))

→ See references/variables.md for complete property reference, replacement strategies (:noquote, :backtick), and usage patterns

Validation

⚠️ MANDATORY for create/update workflows: Validate the dashboard JSON before deploying. Check:

• Schema structure — required top-level keys (name, type, content) and content keys (version, variables, tiles, layouts)
• Variable resolution — all variable queries execute successfully
• Tile query execution — all tile DQL queries run without errors
• Best-practice checks — warnings for hardcoded time filters, CSV variables, etc.

→ Load references/create-update.md for full validation workflow.

References

Common Patterns & Best Practices

Patterns: Executive (header + KPIs + trends) · Service Health (RED metrics) · Infrastructure (resource metrics + tables)

Key rules: Match tile IDs in tiles and layouts · Use descriptive variable IDs · Start with full-width headers (y=0) · Optimize queries with limit/summarize · Set version=21 · No time-range filters in queries unless explicitly requested by the user