v4.70.0

Added

• Added PodsUnschedulable alert

Changed

• LokiObjectStorageLowRate: don't page when WCs are being created

v4.69.0

Added

• add GrafanaPostgresqlRecoveryTestFailed alerting rule.

Changed

• PrometheusOperatorRejectedResources: only page for MC resources

Removed

• DuplicatePrometheusOperatorKubeletService was for clusters before v20, which we don't have anymore.

v4.68.0

Changed

• Update CoreDNS alerts to page only for resources in "kube-system" namespace.
• Route FluxKustomizationFailed for silences kustomization to Atlas.

v4.67.0

Changed

• FluentbitDropRatio only pages for management cluster instances (giantswarm-managed).

Removed

• Removed FluentbitTooManyErrors alerts, at this is already covered by FluentbitDropRatio alerts and they mostly page together.

v4.66.0

Added

• Added cancel_if_metrics_broken inhibition to following alerts:
  • ManagementClusterDeploymentMissingCAPA
  • ManagementClusterDeploymentMissingCAPI
  • ETCDBackupMetricsMissing
  • PrometheusMissingGrafanaCloud
  • MimirToGrafanaCloudExporterDown
  • ManagementClusterDexAppMissing
• Add CiliumAgentPodPending alert for Cabbage.

Changed

• LogForwardingErrors description improvement

v4.65.1

Changed

• Increase MimirIngesterNeedsToBeScaledUp alert's time to trigger from 6h to 12h to avoid noise coming from temporary spikes.
• WorkloadClusterWebhookDurationExceedsTimeoutSolutionEngineers alert: make it page only during business hours, and increase delay to 1h before it pages
• MetricForwardingErrors alert: make it less sensitive

v4.65.0

Changed

• Improved ClusterAutoscalerFailedScaling alert expression to reduce false positives by detecting ongoing scaling failures rather than cumulative historical failures.

v4.64.0

Changed

• Removed grafana from DeploymentNotSatisfiedAtlas because it's already monitored via GrafanaDown alert.
• Rework Rocket's ManagementClusterContainerIsRestartingTooFrequently to use pod names as the selector.
• Update alert for Cilium HelmRelease to match timeout.

v4.63.0

Added

• Add IncorrectResourceUsageData alert.

Changed

• Made MimirIngesterNeedsToBeScaledUp alert less sensitive to CPU usage.
• Increase MimirIngesterNeedsToBeScaledUp alert's time to trigger from 1h to 6h to avoid noise coming from temporary spikes like from stable-testing installations (giantswarm/giantswarm#33513)
• Rewrite Flux alerting rules towards the gotk_resource_info emitted by the Kube State Metrics.
• Drop customer-related alerting rules of Flux.
• Rules unit tests: support for $provider template so we can move provider-specific tests to global tests.
• Rules unit tests: simplify files organization by removing the capi folder. Also fixes a bug in cloud-director tests.
• Rules linting: run against all configured providers.
• Exclude more containers from Rocket's ManagementClusterContainerIsRestartingTooFrequently alert.

v4.62.0

Added

• Add AppAdmissionControllerWebhookDurationExceedsTimeout alert, business hours only.

Removed

• Remove app-admission-controller from generic ManagementClusterWebhookDurationExceedsTimeout alert.

Changed

• Remove duplicate test files for Atlas since all tests are the same accross all CAPI providers.
• Remove duplicate test files for Honeybadger since all tests are the same accross all CAPI providers.
• Remove duplicate test files for Shield since all tests are the same accross all CAPI providers.
• Remove duplicate test files for Tenet since all tests are the same accross all CAPI providers.