Dashboard 2.0 and PromQL

Dashboard 2.0 provides an at-a-glance view of the performance and health of your infrastructure and services. Powered by the Prometheus Query Language (PromQL), you can query metrics for most use cases to quickly identify issues. Dashboard 2.0 permits you to:

• Visualize any data collected using flexible queries.
• Customize charts for personalized data visualization.
• Leverage out-of-the-box content to get metrics from commonly used services.
• Create new dashboards and reuse existing dashboards for easy portability.
• Create data-rich charts using PromQL.

Getting started

Follow simple steps to create the visualization you want with data populated by PromQL query definitions, as shown in the following examples:

Example 1

PromQL query to return multiple time series: sum(usage_cloud_provider_aws)

Visualization:

Example 2

PromQL query to graph the inbound and outbound traffic on an ec2 instance: usage_cloud_provider_aws{instance="ec2.instances"}

Visualization:

Basic create dashboard steps

Open the dashboard menu to:

• Create a dashboard using the dashboard builder tools
• Import a dashboard
• Upload a JSON file that can be used to instantiate a dashboard

To create a dashboard manually:

1. Create the dashboard.
2. Create a metric, text, or image tile.
3. Enter a PromQL query.
4. Select from Combo, Value, or List chart type.
5. Customize the visualization.
6. View the dashboard on the main dashboard page.

References

• OpsRamp Blog: A Dashboard Guide for IT Metrics
• OpsRamp Blog: Defining Killer SLOs with OpsRamp
• Querying Prometheus

Classic Dashboard and Dashboard 2.0 Differences

Current Dashboard 2.0 constraints

Dashboard 2.0 constraints

• Available to auto-monitoring clients, only. At the client level, the dashboard is associated with the user. Partner level can view dashboards in the context of any clients. A client cannot view a dashboard that a partner created.
• There is no public dashboard API.
• Only metric data is handled, not alerts or incidents.
• Only applies to cloud and cloud native resources.
• Dashboard names within a tenant must be unique.
• Curated dashboards can be viewed only for installed integrations.

PromQL constraints

• No range vectors.
• Unsupported functions:
  • sort
  • group by; an only query on individual resources
• Cannot query on resource groups, only individual resources.
• Cannot group on specific values.
• No truncation of decimal values.

Dashboard menu

Dashboard list

The dashboard list shows curated, pre-configured dashboards and user-created dashboards. Click on the dashboard entry to view the dashboard. Dashboards are only displayed if the associated integration is installed.

PromQL query builder

The time series and operations specified in the query power the dashboard visualization. After selecting CREATE TILE, choose the tile type you want and click Data to enter one or more PromQL query definitions.

Use the PromQL language constructs described, below, such as attributes, operators, and functions, to select and aggregate time-series data in real-time and slice time series to create the visualization you want.

Use the dashboard autocomplete feature to show all available matching metrics.

When specifying tags, enter a bracket, {, to get a list of all tags associated with the resource.

In the Legend input field, use double-brackets, {{, to list all available tags for the resource. The legend filters and only shows matching tags for the resource.

You can specify multiple queries in a query definition. Click + to add queries.

General query definition syntax

A PromQL query definition has the following general form:

function(metricName{attribute="value"})

The most basic definition is to specify the tag name to select a time series. For example:

google_storage_network_sent_bytes-count

The PromQL language also supports filtering by attribute, functions, operators and regex expressions:

Visualization

After specifying the metrics you want to populate the dashboard, choose Visualization to create the dashboard visualization.

Metric tiles

Metric tiles support PromQL-driven line and bar time-series, single-value, and list chart types of up to 30 time series.

Chart Type	Name	Description	Example
Single-value	Value	Displays a metric value at one point in time. Queries that return multiple series must have an aggregation. For example, avg(system_cpu_usage_total)
Time-series	Combo	Displays metric values over time and can support up to two queries. The metric can be represented by a line chart or a bar chart.
List	List	Displays the returned series and associated metric values at a single point in time, in list form.

Metric tile customization

Dashboard 2.0 provides a rich toolset for customizing visualizations:

Text tiles

Use text tiles to provide messages and descriptions for dashboard content.

Image tiles

Images can be uploaded as tiles.

Dashboard options

Filters

Collections

Collections group dashboards, providing a way to organize dashboards. You can copy and move dashboards to a collection.