Skip to content
Osuite Osuite Docs

Dashboard builder

Overview

Section titled “Overview”

The dashboard builder is the point-and-click way to create and edit dashboards. It produces the same object as the Visualizer Agent and the dashboards-as-code workflow, so a dashboard started in the UI can later be pulled and edited as JSON, and vice versa.

If you would rather describe the dashboard than build it, the Visualizer Agent writes the PromQL and assembles the panels for you.

Create a dashboard

Section titled “Create a dashboard”

Open the Dashboards section and create a new dashboard. Give it a name; a description and tags are optional.

A new dashboard starts empty. You build it up one panel at a time.

Add a panel

Section titled “Add a panel”

Add a panel to the dashboard. For each panel you set:

  • Title — the panel heading, and an optional description.
  • Queries — one or more PromQL queries that supply the data (see below).
  • Chart type — how the data is drawn: line, area, stacked-area, bar, stacked-bar, scatter, pie, bar-gauge, gauge, stats, table, or service-graph.
  • Unit — how values are formatted: number, percentage, bytes, bytes_per_second, or duration_ms. Pick the unit that matches the metric — bytes for memory, duration_ms for latency, percentage for a utilization ratio.

Attach PromQL queries

Section titled “Attach PromQL queries”

Each panel holds one or more PromQL queries. Use multiple queries when you want to show related series in one panel — for example a request rate and its error rate together. Use separate panels for unrelated metrics.

If you are not comfortable writing PromQL by hand, the Write PromQL Agent composes a query from a plain-English description against your real metrics.

Choose a chart type

Section titled “Choose a chart type”

Match the chart type to the shape of the data:

  • Continuous measurements over time → line or area.
  • Series that sum to a whole → stacked-area or stacked-bar.
  • A single current value → gauge, bar-gauge, or stats.
  • Proportions at a point in time → pie.
  • Raw rows → table.
  • Service topology → service-graph.

The full list and what each is for is in the dashboards overview.

Arrange panels

Section titled “Arrange panels”

Panels are laid out on a 24-column grid. A full-width panel spans 24 columns; a half-width panel spans 12. Drag to move a panel and drag its edge to resize it.

Use variables

Section titled “Use variables”

To make one dashboard work across many services, namespaces, or environments, add template variables and reference them in your queries as $name (for example $service or $namespace). The dashboard then exposes selectors at the top that repoint every panel at once.

Save and publish

Section titled “Save and publish”

Save the dashboard to persist it. Publishing a dashboard makes it visible beyond your own editing view.

Next steps

Section titled “Next steps”