The new alerts in Grafana 8.0 are an opt-in feature that centralizes alerting information for Grafana managed alerts and alerts from Prometheus-compatible data sources in one UI and API. You can create and edit alerting rules for Grafana managed alerts, Mimir alerts, and Loki alerts as well as see alerting information from prometheus-compatible data sources in a single, searchable view. For more information, on how to create and edit alerts and notifications, refer to Grafana Alerting.

As part of the new alert changes, we have introduced a new data source, Alertmanager, which includes built-in support for Prometheus Alertmanager. It is presently in alpha and it not accessible unless alpha plugins are enabled in Grafana settings. For more information, refer to Alertmanager data source.

To learn more about the differences between new alerts and the legacy alerts, refer to What’s New with Grafana 8 Alerts.

Library panels allow users to build panels that can be used in multiple dashboards. Any updates made to that shared panel will then automatically be applied to all the dashboards that have that panel. For instructions on how to create, add, unlink and manage library panels, refer to Library panels.

Data sources can now send real-time updates to dashboards over a websocket connection. This can be used with the MQTT data source.

In addition to data source integration, events can be sent to dashboards by posting metrics to the new live endpoint: /api/live/push endpoint.

These metrics will be broadcast to all dashboards connected to that stream endpoint.

For more information about real-time streaming, refer to Grafana Live documentation.

The Prometheus metrics browser allows you to quickly find metrics and select relevant labels to build basic queries. If supported by your Prometheus instance, each metric will show its HELP and TYPE as a tooltip. For more information, refer to Metrics browser documentation.

The Bar chart panel is a new visualization that supports categorical data. It only supports one data frame and it needs to have at least one string field that will be used as the category for an X or Y axis and one or more numerical fields.

To use it with time series you first have to add a Reduce transform.

For more information, refer to Bar chart visualization.

The State timeline visualization shows discrete state changes over time. Each field or series is rendered as a unique horizontal band. This panel works well with string or boolean states, but it can also be used with time series data. When used with time series data, the thresholds are used to turn the numerical values into discrete state regions.

This panel also takes advantage of the new value mapping features that allow you to color string and boolean values.

Example with string values:

With time series data and thresholds:

For more information, refer to State timeline visualization.

A sister panel to the state timeline is the new Status history panel visualization. It can display periodic states in a grid view. It supports both numerical, string, or boolean states. You can assign colors using value mappings, thresholds, or gradient color. For more information, refer to Status history. schemes.

This hidden feature of the old Graph panel is now a standalone visualization. It combines a histogram transformation and bar chart visualization into a single, integrated, easy-to-use panel. There is also a new standalone histogram transformation that can be paired with any visualization.

image::histogram-8-0.png

For more information, refer to Histogram

The Time series is out of beta! We are removing the Beta tag and graduating the Time series visualization to a stable state.

Time series panel topics have been updated as a result of these changes.

You can now expand the node graph for the displayed trace when using the Trace to logs feature. Depending on the data source, this can show spans of the trace as nodes in the graph, or as some additional context like service map based on the current trace.

We also added a grid view and the ability to explore hidden nodes.

Tracing in Explore and Node graph were updated as a result of these changes.

The Pie chart is out of beta! We are removing the beta tag and graduating the Pie chart visualization to a stable state.

Lots of panel editor improvements, heavily informed by user research and community feedback.

The Panels section has been updated to reflect these changes.

Grafana 8 comes with a refreshed look and feel, including themes changed to be more accessible. The improved Grafana UI brings a number of adjustments and tweaks that make the application even more fun to use. Enjoy the new home dashboard design!

Under the hood, the new theme architecture enables us to bring more sophisticated themes control in the future.

When you inspect a panel, you can now download log results as a text (.txt) file.

Download log results was added as a result of this feature.

The new Explore inspector helps you understand and troubleshoot your queries. You can inspect the raw data, export that data to a comma-separated values (CSV) file, export log results in text format, and view query requests.

Inspector in Explore was added as a result of this feature.

Log navigation in Explore has been significantly improved. We added pagination to logs, so you can click through older or newer logs as needed.

Logs in Explore was updated as a result of these changes.

image::navigate-logs-8-0.png

You can now use the Plugin catalog app to easily manage your plugins from within Grafana. Install, update, and uninstall plugins without requiring a server restart.

Plugin catalog was added as a result of this feature.

Grafana 8.0 includes many performance enhancements.

The following data source updates are included with this Grafana release.

This Grafana release includes the following authentication updates.

Grafana has updated its license from Apache 2.0 to the GNU Affero General Public License (AGPL). Please see the related blog post, Q&A and license for more details.

You can now add or remove detailed permissions from Viewer, Editor, and Admin org roles, to grant users just the right amount of access within Grafana. Available permissions include the ability to view and manage Users, Reports, and the Access Control API itself. Grafana will support more and more permissions over the coming months.

Role-based access control docs were added as a result of this feature.

Grafana caches the results of backend data source queries so that multiple users viewing the same dashboard or panel do not make multiple submissions of the same query to the data source (like Splunk or Snowflake) itself.

This results in faster average load times for dashboards and fewer duplicate queries overall to data sources, which reduces cost and the risk of throttling, reaching API limits, or overloading your data sources.

You can enable caching per data source, and time-to-live (TTL) can be configured globally and per data source. Query caching can be set up with Redis, Memcached, or a simple in-memory cache.

For more information, refer to the Data source query caching docs.

When creating a report, you can now choose to export Table panels as .csv files attached to your report email. This makes it easier for recipients to view and work with that data.

You can also link back to the dashboard directly from the email, for users who want to see the data live in Grafana. This release also includes some improvements to the Reports list view.

For more information, refer to Reporting docs.

The Grafana Enterprise documentation has been updated to describe more specifically how licensed roles are counted, how they can be updated, and where you can see details about dashboard and folder permissions that affect users’ licensed roles.

For more information, refer to License restrictions docs.

Documentation was updated to reflect these changes.

For multi-search requests, we now use the correct application/x-ndjson content type instead of the incorrect application/json. Although this should be transparent to most of the users, if you are running Elasticsearch behind a proxy, then be sure that your proxy correctly handles requests with this content type.

The default HTTP method for Prometheus data source is now POST, previously it was GET. The POST APIs are there since January 2018 (Prometheus 2.1.0) and they have fewer limitations than the GET APIs. Users with Prometheus instance with version < 2.1.0 that use the default HTTP method should update their HTTP method to GET.