Agent0 generates dashboards from natural language descriptions. Describe what you want to visualize, and Agent0 creates the dashboard with panels, queries, and visualizations based on your actual telemetry.

How It Works

Agent0 analyzes your natural language description and automatically:

Generates valid PromQL queries based on your telemetry
Selects appropriate visualization types (time series, pie chart, gauge, etc.)
Configures panel layouts and groupings
Sets meaningful panel titles and descriptions
Validates that the underlying queries return real data before creating the dashboard

No broken panels are emitted — if a query doesn't resolve, Agent0 reports the failure rather than creating a non-functional dashboard.

Open Agent0

Access Agent0 from:

Agent0 Chat page — https://app.dash0.com/agent0
Command palette — Press ⌘K (Mac) or Ctrl+K (Windows/Linux) from any page
In-page buttons — Look for Agent0 buttons in explorers and chart views

The Integrated Experience button appears contextually throughout Dash0, such as in the Service Map shown above. When you click this button, Agent0 understands your current context and can help create dashboards based on what you're viewing.

Select Dashboards & alerts → Creating a dashboard to start with a focused prompt.

Describe What You Want to Visualize

Be specific about:

The services or components you want to monitor
The metrics you want to track (latency, errors, throughput, resource usage, etc.)
The time range or granularity
Optional: visualization preferences (charts, gauges, tables)

Example prompts:

"Create a dashboard showing RED metrics for the frontend service"
"Build a dashboard tracking database performance with connection pool, query latency, and error rates"
"Show me a dashboard for Kubernetes pod health with CPU, memory, and restart counts"
"Create a service overview dashboard with latency, error rate, and throughput for all services"
"Build a dashboard comparing p50, p95, and p99 latency across microservices"
"Show infrastructure metrics for the payment service: CPU, memory, network, and disk I/O"

Tip

The more specific your prompt, the better Agent0 can configure the dashboard. Include service names, metric types, and visualization preferences for optimal results.

Review the Generated Dashboard

Agent0 generates a complete dashboard configuration:

Panels — Individual visualizations with queries and display settings
Layout — Panel positioning and sizing
Queries — PromQL expressions validated against your telemetry
Visualizations — Appropriate chart types for each metric
Variables — Dashboard-level filters (e.g., service selector, environment)

Example for "Create a dashboard showing RED metrics for the frontend service":

Panel 1: Request Rate — Time series chart showing requests per second
Panel 2: Error Rate — Time series chart showing error percentage
Panel 3: Latency — Time series chart showing p50, p95, p99 latency
Variables — Service selector defaulting to "frontend"

Refine the Configuration

If the generated dashboard needs adjustments, iterate with Agent0:

"Add a panel showing top 10 endpoints by latency"
"Change the latency panel to show only p99"
"Add memory and CPU usage panels"
"Group panels by service instead of metric type"
"Add a variable to filter by environment"

Save the Dashboard

Once satisfied with the configuration, click View dashboard in the Agent0 response.

The dashboard opens in edit mode with all panels pre-configured. Review the layout and visualizations, then click Save dashboard to persist it.

The dashboard is immediately available in your dashboard list and can be shared with teammates through access control permissions.

Generate Infrastructure-as-Code

Agent0 can also generate the corresponding Terraform configuration for the dashboard and raise it as a pull request to your infrastructure repository, keeping your observability configuration in code.

See Manage Dashboards as Code for details on version controlling dashboard configurations.

Generate Dashboards with Agent0