Changes and Releases
Updates on Giant Swarm workload cluster releases, apps, UI improvements and documentation changes.
Changed
- Upgrade grafana chart: 11.2.3 => 11.3.3
- Upgrade grafana (appVersion): 12.4.0 => 12.4.1
Added
- Add deployment edit mode with DeploymentPicker and EntityPicker scaffolder fields.
- Add MCP tool get-helm-chart-values for fetching Helm chart default values and schema.
- Add ConfigMap and Secret resource classes, HelmRelease valuesFrom inspection methods, and core API resource path support.
- Dereference external schema references in app deployment configuration.
Changed
- Reduce system prompt instructions for muster.
Fixed
Changed
- Upgrade Tempo chart from to 2.4.2 to 2.6.2
- Upgrades Tempo from 2.10.1 to 2.10.2
Added
- requests and limits for distributor - fixes its HPA
Fixed
- vulture search window reduced to 24h to avoid querying traces out of retention period
Fixed
- Fix Azure storage management policy rule YAML structure for lifecycle policy.
Added
- Add workload status to deployment details page.
Fixed
Added
- Add
httpRouteFilter support to Gateway API routes, enabling HTTPRouteFilter resources for tenant header enforcement (X-Scope-OrgID). - Add
backendTrafficPolicy support to Gateway API routes. Uses a free-form spec object to allow any Envoy Gateway BackendTrafficPolicy fields (timeout, requestBuffer, circuitBreaker, etc.) rather than hardcoding a single field.
Added
- Pre-configure
httpRouteFilter.directResponse default (400 Bad Request) so that enabling httpRouteFilter returns a meaningful response without requiring callers to specify the response body.
Added
- Add AWS account ID to cluster details page.
- Add GiantSwarmLocationProcessor to handle “giantswarm” catalog location type, assigning the giantswarm namespace to entities from GS-sourced locations.
- Add custom session header support for MCP servers that use non-standard session headers.
Fixed
Changed
- Chart: Sync to upstream.
- Controller: Update image to v1.15.0.
Changed
- Enable priority class by default to ensure fluentbit can preempt lower priority pods on overloaded nodes.