The functionality in this topic is for unified dashboards and not dashboards for Streams.
You can monitor both your metric and span data by adding charts to unified dashboards based on queries to your telemetry data. Dashboards can contain any number of charts from both metric and span data.
Cloud Observability offers a number of pre-built dashboards you can use without having to query your data. Start with one of these then customize it to fit your needs. You can edit the description, labels, and sections, add/delete visualizations, and change chart queries.
Cloud Observability provides two types of pre-built service dashboards. Both provide an overview of performance for a service, its operations, and if you’re ingesting Kubernetes metrics, the associated metrics that the service is emitting.
Seeing the transaction level information about your service (SLIs from spans) and infrastructure health (metrics) in the same dashboard provides a powerful view into the overall health of your system.
Global service dashboard: A single dashboard that uses a template variable for the service, allowing you to view all your services from one dashboard. Change the
service variable value to view specific services.
Individual service dashboard: Available for every service reporting to Cloud Observability. This dashboard provides the same information but because it’s scoped to a single service, it also provides an additional section about the performance of the individual operations on that service.
Both types of service dashboards use the
span_kind variable to determine the data to display. You can choose from the following values:
server(default): Shows data from spans providing functionality asked for by other services. Use this value for back-end services whose performance may affect other services in the request stream.
client: Shows data from spans requesting something from another service. Use this value when other services rely on these requests, such as the entry point to your system.
internal: Shows data from spans that are completely internal to the service.
all: Shows data from all spans from the service.
Using Terraform? You can use the Cloud Observability Terraform provider to create and manage your dashboards and charts. You can also use it to export existing dashboards into the Terraform format.
From the navigation bar, click Dashboards.
Click Create a pre-built dashboard and choose a dashboard.
Pre-built dashboards are not available for all AWS, Azure, and Collector integrations.
This image shows how to create a dashboard. The steps below describe the procedure in more detail.
From the navigation bar, click Dashboards.
Click Create a custom dashboard to access a new, blank dashboard.
Click Details and in the Details panel, rename the dashboard and optionally add a description and labels.
The description appears on the dashboard and Dashboard List view. The description can be up to 256 characters, and it supports Markdown. Here’s an example:
1 2 3 4 5 6 7 8 Use this dashboard to monitor and identify issues with Cloud Observability's customer experience. Have questions? Ask in [#customer-experience](https://lightstep.com/). Docs: - [Runbooks](https://lightstep.com/) - [Troubleshooting](https://lightstep.com/)
If you’re description needs more than one line, consider adding a text panel instead.
You’re now ready to add charts to the dashboard.
You can also add a chart to the dashboard from the Service Directory view.
Once you create a dashboard, you can edit it by adding more charts, creating sections, and adding text blocks. You can change the display of the charts by setting time periods and using template variables.
Be careful when opening the same dashboard in more than one tab or window. Edits may not carry over to multiple instances.
Add a chart by clicking Chart and then follow these instructions to create your chart.
Along with charts that visualize your data, you can add text panels to display content such as a description of the section and links to related content.
Click Text panel and then add Markdown to the panel.
You can organize your dashboard by adding named sections.
This image shows how to add a section. The steps below describe the procedure in more detail.
At the top of the page, click Section. The new section is placed at the bottom of the dashboard.
By default, the section is named New Section. Click into the title to rename it.
The first blank chart is added for you. Click on it to create your query and configure you chart.
To add more charts, click Add > Chart (or Text panel). To rearrange charts, click the title of a chart to drag and drop it into a section.
To move a section, click the More ( ⋮ ) icon to move it up or down the dashboard. You can also expand and collapse sections from here.
When you delete a section, you can choose to delete just the section (the charts remain in the same place in the dashboard), or you can remove the section and its charts.
You can move charts added to a dashboard by dragging and dropping them. Hover over a chart title. When the cursor turns into a hand, drag and drop the chart into its desired location.
You can resize a chart by hovering over the lower right-hand corner and dragging to the size you need.
You can add labels (like
service:myservice) to a dashboard.
Learn how to add labels.
Template variables allow you to dynamically filter charts in a dashboard based on attribute values returned by the data. Instead of creating a chart for each value, you create one chart and then filter the dashboard to specific values.
For example, say you want to view the health of your services. Instead of creating a separate dashboard for each service, you create one dashboard that uses
service as a template variable. On initial load, the charts show data using the default value of the variable (usually
all, so in this case, from all services). You use the variable dropdown to select only the services you’re interested in.
Template variables provide greater functionality and are more predictable than Global filters. You can choose to use either one at the top of the dashboard.
To learn more, see Use template variables.
If you want to quickly view data filtered to an attribute and don’t need to persist that ability, you can change the data reported on the dashboard using global filters.
Global filters only affect your view - they are not persisted.
To use global filters, click the More ( ⋮ ) icon and choose Switch to global filters
When adding multiple filters, Cloud Observability narrows down your choices based on
attributes available after the previous filter. Multiple selections use
When you add a global filter, you’re changing the query for each chart on the dashboard. If a chart is using a Stream to retain a span query and you apply a global filter, it’s considered a new query that is not retained (you will only see data currently in the retention window). If you add a chart with a Stream’s retained query to a dashboard that uses global filters, the query is retained as a Stream without those filters applied.
By default, Cloud Observability shows data from the past 60 minutes. To change the time range, click the time-picker drop-down and:
Enter custom time ranges in the From and To inputs.
The time picker saves custom time ranges and shows them under Recently used. You can also click < > next to the time picker drop-down to move back and forward in time.
To view live data for the last 60 minutes, click Play next to the time picker. When viewing live data, you can do the following:
To stop live viewing live data, click Pause. Cloud Observability stops refreshing the chart and sets a fixed, custom time window matching the last refresh.
Changing the time period only affects your view - the change isn’t persisted.
Data may not be available for the time range you select, based on the type of data and your data retention policy.
There are two ways to clone an entire dashboard:
In the dashboard, click the More ( ⋮ ) icon and select Clone.
In the Dashboard List view, click the More ( ⋮ ) icon next to the relevant dashboard and select Clone.
Cloned dashboards do not copy over the global filters and they reset to the default time period of the last 60 minutes.
Follow these steps to clone a chart on a dashboard:
Cloned charts use any global filters and the time period set for the dashboard.
Follow these steps to delete a chart:
Deleting a chart from a dashboard deletes it from everyone’s view of the dashboard.
You can delete a dashboard from either the dashboard itself or from the Dashboard list. Deleting a dashboard also deletes all charts created for it.
To delete it from the dashboard, click the More ( ⋮ ) icon, select Delete, and click Yes, delete in the dialog.
Cloud Observability creates a short link that you can use to share a dashboard. Any filters or time ranges applied to the dashboard are maintained when the other person visits the link.
To get the short link, click the More ( ⋮ ) icon and select Share. The link is copied to your clipboard.
You can jump to another dashboard using the dropdown next to the dashboard title.
If you use Terraform to manage your dashboards, a common use case is to create a “baseline” dashboard in the Cloud Observability UI and then export it to Terraform to create and manage new dashboards. Creating a dashboard in the UI is easier and more intuitive than working in a Terraform configuration file. When you export the dashboard to a Terraform file, you can then clone it and edit the file to create new dashboards, for example by changing the individual queries behind the charts.
To export a dashboard to Terraform, click the More ( ⋮ ) icon and
choose Export to Terraform. The
tf config file can then be downloaded to
any local directory.
To get a better look at a chart on a dashboard, you can enlarge it. Hover over the chart and click the Expand chart icon (CTRL+v).
The chart opens in the editing view with the query builder collapsed. Click Query to open the builder and edit the query.
If the chart contains multiple queries or queries and a formula, you can toggle their display on and off below the chart.
Below the chart, a table displays the data for each line in the chart.
The toggle selection is not persisted to the dashboard.
You manage all your dashboards from the Dashboard view. By default, dashboards are listed in alphabetical order.
Sorting and dashboard favorites are saved only for a user. Other users will not see your sort or favorites.
You can sort the dashboards by any column.
Use the Search by field to find a dashboard. You can search by name, date, type, or creator’s name.
To search by a label, use the Search by labels field.
On the Dashboard list page, click the star to favorite a dashboard. Favorited dashboards always appear at the top of the dashboard list.
Select the dashboards to delete and click Delete selected dashboards.
You can create an alert from a chart, but because alerts require a specific structure, not all charts can be used to create one. If your query contains a group-by, you will not be able to create an alert (the option is disabled). For queries that don’t meet other requirements, you’ll be able to edit the query on the Alert Configuration page.
Creating an alert from a span query automatically creates a Stream.
Follow these steps to create an alert from a chart:
Hover over the chart in the dashboard, click the More ( ⋮ ) icon, and select Create an alert.
The Alert Configuration page opens in a new tab using the query from the original chart. A banner describes the edits needed to create a valid alert. Fields in violation are highlighted.
Fix the violations.
By default, the title is the same as the original chart. You can change it and you can add a description, if needed.
Continue creating the alert.
You can add to a notebook for when, during an investigation, you want to be able to run ad hoc queries, take notes, and save your analysis for use in postmortems or runbooks. Notebooks allow you to view metric and trace data from different places in Cloud Observability together, in one place.
To add to a notebook, click Add to notebook and search to choose an existing notebook or create a new notebook.
When you add to a notebook, a chart is created using the same query. The annotation is a link back to the original, so you can quickly return to the origin of your investigation.
Learn more about notebooks.
When you notice issues on dashboards, use Cloud Observability’s correlation feature to investigate the deviation and find possible causes.
You can’t use the correlation feature on big number charts.
To run the correlation feature, click View correlations or click directly in the chart and select View correlations. Cloud Observability opens a side panel where you can begin your investigation.
Visit Investigate deviations for more information.
Updated Sep 21, 2023