Discover Metrics

The Metrics page in OpenSearch Dashboards provides a dedicated interface for discovering, querying, and visualizing time-series metric data. It is optimized for working with Prometheus metrics using PromQL.

The page has two modes, switched with the radio toggle at the top:

• Explore — a no-code metric browser for finding metrics and filtering by labels.
• Query — a PromQL editor for authoring and visualizing expressions.

Both modes share the same time range picker and data source selector; switch at any time.

The Metrics page is available in Observability workspaces. To access it, navigate to an Observability workspace, then in the left navigation expand Discover and select Metrics.

Explore mode

Browse and discover metrics without writing PromQL. Use it to find a metric when you don’t yet know what to query.

The interface includes:

• A metric browser that lists all metrics from the data source as cards, each with a small sparkline preview.
• A search box with debounced server-side filtering by metric name.
• Label filters that narrow the metric list to only those carrying matching labels. Each filter row uses a Prometheus operator: =, !=, =~ (regex match), or !~ (regex not match).
• Grouping options that organize metric cards by Prefix (the namespace prefix of the metric name) or A-Z (alphabetical, in a single group).
• Layout options for Grid or Rows.
• Metric type badges — Counter, Gauge, Histogram, Summary, or Unknown — taken from Prometheus metric metadata, with a fallback that infers Counter and Histogram from name suffixes (_total, _count, _sum, _created, _bucket) when metadata is missing.

Click a metric card to open the metric detail view, which previews the time series broken down by label and lets you drill into a specific label combination. Or select one or more metric cards and select Query Metrics to send them into Query mode as a multi-query execution.

Query mode

Author one or more PromQL expressions and render them in a chart. Use it when you know what you want to compute.

Each query lives in its own row. A single execution can run several queries together — they’re concatenated with semicolons (;) and rendered side by side on the visualization.

Builder vs. Code

Each query row has two editor styles, toggled per row:

• Builder — a structured UI for selecting a metric, adding label filters with operators (=, !=, =~, !~), and chaining operations (aggregations, range functions, math). Useful when you’re learning PromQL or want fewer typos.
• Code — a raw PromQL editor with autocomplete and syntax highlighting. Suggestions are fetched from the Prometheus data source and trigger on space, (, {, [, ,, =, ~, ', and ".

You can switch between Builder and Code at any time. Switching from Builder to Code reveals the underlying PromQL; the reverse parses your PromQL back into Builder controls when possible.

Writing queries

Write queries using PromQL syntax. For example:

up{job="prometheus"}

Multiple queries in one execution

To run several queries together, add a row for each one (or, in Code mode, separate them with semicolons on different lines):

up{job="prometheus"};
node_cpu_seconds_total{mode="idle"};

Each query runs independently. The results are returned together and the visualization renders all series side by side.

Step and resolution

In Explore mode, sparkline and detail queries are generated client-side. The step (data resolution, in seconds) is derived from the active time range — targeting roughly 1440 datapoints with a 15-second floor — and the rate() window is sized to be at least four times the assumed 60-second scrape interval and at least one scrape interval larger than the step, so zoomed-out views don’t show gaps.

Configuring a Prometheus data source

Before you start, configure a Prometheus data source using one of the following methods.

Configuring a Prometheus data source in OpenSearch Dashboards

To configure a Prometheus data source in OpenSearch Dashboards, follow these steps:

1. In the left navigation, go to Data Administration > Data sources.
2. Select Create data source.
3. Select Prometheus.
4. Enter a Data source name and optional Description.
5. Enter the Prometheus URI endpoint (for example, http://prometheus-server:9090).
6. Configure the Authentication method:
   • No authentication: Use if your Prometheus server does not require authentication.
   • Basic authentication: Enter a username and password.
   • AWS Signature Version 4: Use for Amazon Managed Service for Prometheus.
7. Select Connect.

Configuring a Prometheus data source using the API

Alternatively, you can configure a Prometheus data source programmatically. For more information, see the Data sources documentation.

Time filter

Use the time filter to specify the time range for your metric data:

• Quick select: Choose a relative time range (for example, the last 15 minutes or the last 1 hour).
• Commonly used: Select from predefined time ranges.
• Custom: Specify absolute start and end times.
• Auto-refresh: Set an automatic refresh interval.

Viewing results

After running a query in Query mode, the results are displayed in a tabbed interface below the editor containing the following views:

• The Visualization tab provides interactive charts for your metric data.
• The Table tab displays each series as a row of latest values, with a banner if the response was truncated for performance.

Series limit

To keep the browser responsive, the visualization renders up to 20 series by default. If a query returns more, a banner appears with a Show {count} button to render all of them. Tighten label filters or aggregate (for example, sum by (...)) to bring the count down rather than rendering hundreds of series.

Configuring visualizations

When the Visualization tab is selected, a settings panel appears on the right side of the screen. Use this panel to:

1. Select a chart type: Choose from line, bar, area, scatter, pie, heatmap, table, metric, gauge, bar gauge, state timeline, or histogram.
2. Map axes: Assign fields to the X and Y axes.
3. Customize styles: Adjust colors, legends, gridlines, thresholds, and units.

When you modify the settings, the visualization is updated automatically. See Build a Dashboard for details on each chart type and its options, and Visualization transformations for reshaping query results before they’re charted.

Saving and reusing

Metrics explorations are saved as Explore saved objects. Save a query from the page actions to revisit it later, or add a metrics view as a panel on a new or existing dashboard to track it alongside logs and traces. From the global Visualize application, choosing Create visualization → Visualize with Metrics also opens a fresh metrics canvas in Query mode.