What's new from Grafana Labs

Remote configuration is a core feature of Grafana Fleet Management. As the number of configuration pipelines grows, browsing and managing them can become more cumbersome when navigating through paginated lists.

Grafana Fleet Management has added a Changed by column to the pipeline history view. Now you can see who made the change as part of your audit trail.

SQL expressions are now generally available. SQL expressions let you query, join, and reshape the results of your existing data source queries using familiar MySQL-like syntax. Instead of chaining multiple transformations to get your data into the right shape, you can write a single SQL statement that filters, aggregates, and combines results from one or more queries. This is especially powerful when you need to:

Locate your erroring collectors faster with new filter functionality from the Overview tab in Grafana Fleet Management.

Click on a status section of the Collector health status chart to be taken to a pre-filtered version of your Inventory.

We redesigned the query variable editor with a new editing experience that makes working with query variables easier.

The new experience uses a dialog box with the values preview at the top and a tabbed editor below, so you can review results while you edit. The wider dialog box also provides plenty of space to work without so that you don’t have to do as much scrolling to see all the query settings.

Grafana Fleet Management now supports OpenTelemetry Collectors, so you can monitor and remotely configure Grafana Alloy instances and upstream OpenTelemetry Collectors from a single control plane.

You can now select which k6 major version your script targets when you create a test in the script editor. New tests default to v1.x (default). Select v2.x when your script uses the k6 v2.x runtime. Grafana Cloud k6 stores the version on the load test and uses it for future runs, including scheduled runs.

Troubleshoot your entity setup with Knowledge Graph mode

Knowledge Graph mode is a new specialist mode in Grafana Assistant that understands how your entity graph is built, end to end, and helps you work with it: ask why it looks the way it does, fix what doesn’t look right, and add what doesn’t ship by default.

Storing a service account JSON key file to connect Grafana Cloud to Google Cloud is a persistent security liability. Keys get embedded in config files, shared across teams, or quietly forgotten — all while remaining valid for years. When one leaks, there’s no automatic expiry and no limited blast radius.

The BigQuery and Google Cloud Monitoring data source plugins support Google Cloud Workload Identity Federation (WIF) as an authentication method on Grafana Cloud. Instead of uploading a key file, Grafana Cloud exchanges the signed-in user’s OIDC token for a short-lived Google Cloud access token via the Security Token Service. Credentials are scoped to the active session and expire automatically — no keys to store, rotate, or accidentally expose.

You can now star test runs in Grafana Cloud k6 to mark the results that matter and find them again quickly. Starring gives you one consistent way to flag a test run as a reference point, whether you’re tracking a release, validating a performance optimization, or keeping an important run on hand to compare against.

Creating browser tests is now faster and easier with the new experimental Agentic testing experience in Grafana Cloud.

Describe the user journey you want to test, such as signing in, completing checkout, or submitting a form, and Grafana Cloud helps turn that journey into a runnable browser test you can run, review, and iterate on without manually scripting every step.

In Grafana Cloud, the internal Grafana Alertmanager handles Grafana-managed alerts by default.

• New stacks use internal routing by default. Sending to an external Alertmanager isn’t enabled by default.
• Existing stacks move to internal routing by default, plan by plan. If your stack actively sends to an external Alertmanager, it keeps that configuration and continues to work.

Sending Grafana-managed alerts to an external Alertmanager is still supported. If you need it, contact Grafana Support.

Grafana Cloud Synthetic Monitoring now lets you organize your checks into Grafana Folders. As you create more checks or onboard new teams, a single flat list becomes hard to scan. With folders, you can group checks by team, service, or environment so the right checks are always easy to find and act on a whole group at once instead of one check at a time.

Want more control than our pattern-based recommendations, but not sure where to start?

Adaptive Logs now walks you through creating a drop rule tailored to your environment in three guided steps: pick one of your highest-volume services, choose a drop rate, and review a pre-filled rule before any logs are dropped. The flow targets debug logs, often the largest and least-queried part of your log volume, and you take it from there.

Starting with Grafana 13.1, Grafana Assistant comes pre-installed in Grafana Enterprise, with no plugin installation required. If you’re a Grafana Enterprise user, simply connect your Grafana Cloud account to start using Assistant right away. Get onboarded quicker, explore telemetry data, build dashboards, and get things done faster using natural language.

This builds on our vision of Assistant as an “anywhere agent” and our ongoing efforts to expand access to our purpose-built LLM, including to self-managed Grafana users.