What grafana/grafana shipped
Written by FoxPlug from public releases; not affiliated with Grafana. An automatic summary of the public release, pull request and commit data of github.com/grafana/grafana. Grafana did not write it and does not use or endorse FoxPlug. Every line links to the public change it describes.
Get a weekly update like this for your product, free
Week of September 21, 2026
What shipped
- Router fixes the preferred API version selection and prunes unnecessary OpenAPI documentation. Pull request #133646
- Router's aggregated discovery now caches results instead of calling every backend on each request. Pull request #133588
- Router exposes state and metrics through dashboards to show which backends serve each API group and their health. Pull request #133638
- Router watch requests now behave consistently with Kubernetes, enabling informers and plugin operators. Pull request #133630
- Router bounds response header wait times and classifies proxy failures for better debugging. Pull request #133627
- PrometheusWriter can now split large recording-rule writes into batches to avoid timeouts. Pull request #133299
- SMTP settings are now read live from the config provider, allowing remote overrides without restart. Pull request #133495
- Notebooks use the creation date as part of the default name in both UI and API. Pull request #133470
- Added experimental feature toggle for unified search on saved queries. Pull request #133592
- API builders now expose full manifests with per-kind settings instead of just resource identity. Pull request #133479
Why it matters
The router improvements make it more production-ready by fixing discovery performance, watch compatibility, and adding observability. Configuration changes now take effect without restarts. API improvements enable better integrations with plugin operators.
Changelog entry
- Router: Fix preferred version selection and prune OpenAPI docs Pull request #133646
- Router: Cache aggregated discovery instead of calling backends on each request Pull request #133588
- Router: Expose route state and metrics for dashboard monitoring Pull request #133638
- Router: Make watch requests behave like Kubernetes for informer support Pull request #133630
- Router: Bound response header timeouts and improve failure classification Pull request #133627
- Alerting: Split large recording-rule writes into batches Pull request #133299
- Notifications: Read SMTP settings live from config provider Pull request #133495
- Notebooks: Use date in default notebook name Pull request #133470
- Search: Add experimental saved-query search feature toggle Pull request #133592
- API: Expose full manifests from builders with per-kind settings Pull request #133479
Router discovery now caches results, watches work like Kubernetes, and state is visible through dashboards. SMTP settings apply live from config overrides. Notebooks use creation dates in default names.
This week brought significant router stability improvements: aggregated discovery caching eliminates per-backend calls, watch requests now match Kubernetes behavior for informer compatibility, and dashboards can monitor backend health and API group routing. Configuration is more dynamic with SMTP settings reading live from the config provider. API builders expose richer manifest details for enterprise features.
Week of September 14, 2026
What shipped
- Alerting now supports Amazon Managed Prometheus recording rule writes, fixing failures when the write target is an AMP data source. Pull request #128090
- Alerting widened missing-series resolution counts to BIGINT on MySQL and PostgreSQL to support evaluations above the SMALLINT limit. Pull request #133023
- Dashboards can now render Assistant-proposed dashboard plans as static previews before building them. Pull request #132761
- Notebooks added IRM incident creation and attachment entry points in the overflow menu. Pull request #132280
- Alerting added targetDatasourceUid, persisted dedup hash, and per-organization poll interval for external ruler sync. Pull request #132525
- Storage LIST operations now preserve pagination continuation tokens after authorization filtering. Pull request #133017
- Provisioning fixed the folder picker when moving dashboards or folders managed by folderless Git Sync repositories. Pull request #132453
- Dashboards migrated query editor and runners from deprecated sync DataSource APIs to async replacements. Pull request #131315
- Table panels gained a zebra striping option behind table.refreshNewFeatures using theme tokens. Pull request #132423
- Table panels now fade in a scroll shadow gradient at edges when rows are out of view. Pull request #132579
Why it matters
This week saw improvements across alerting infrastructure with better data source support and larger evaluation counts, storage refinements for authorization and pagination, and dashboard/notebook features for planning and incident management. Query execution migrated to modern async APIs and provisioning workflows improved for Git Sync operations.
Changelog entry
- Alerting: Support Amazon Managed Prometheus recording rule writes Pull request #128090
- Alerting: Support missing-series resolution counts above the SMALLINT limit Pull request #133023
- Dashboards: Static preview for Assistant dashboard plans Pull request #132761
- Notebooks: IRM incident creation and attachment Pull request #132280
- Alerting: Add targetDatasourceUid, persisted dedup hash, and per-org poll interval Pull request #132525
- Storage: Preserve list pagination after authorization filtering Pull request #133017
- Provisioning: Fix folderless Git Sync move picker Pull request #132453
- Dashboards: Migrate query editor and runners to async DataSource APIs Pull request #131315
- Table: Zebra striping option Pull request #132423
- Table: Scroll shadow fade-in at edges with hidden rows Pull request #132579
This week: Alerting now supports Amazon Managed Prometheus, dashboard assistant plans render as previews, notebooks gained IRM incident actions, and storage pagination works better after authorization filtering.
New in Grafana: Amazon Managed Prometheus recording rules now work with Alerting, dashboard assistant plans preview before building, notebooks can create and attach IRM incidents, and storage pagination handles authorization filtering correctly. Query execution migrated to async APIs and Git Sync folder picker improved for folderless repositories.