Helm chart that provides standardized PriorityClass resources for Giant Swarm clusters.
This chart deploys two PriorityClass resources as defined by sig-architecture:
giantswarm-critical(value: 1000000000) - For components that the cluster absolutely requires to run and Giant Swarm is adding (e.g., kyverno)giantswarm-high(value: 900000000) - For important components that should preempt customer workloads (e.g., observability components, crossplane)
These priority classes replace the previous inconsistent per-app priority classes (like crossplane-critical, flux-giantswarm-flux-giantswarm, prometheus) to ensure consistency across all clusters.
Note: system-cluster-critical and system-node-critical are built-in Kubernetes priority classes and are not included in this chart.
system-node-critical(2000001000) - Kubernetes built-insystem-cluster-critical(2000000000) - Kubernetes built-ingiantswarm-critical(1000000000) - This chartgiantswarm-high(900000000) - This chart
See roadmap issue #3483 for the full context and sig-architecture decision.
This chart is automatically deployed to workload clusters via HelmRelease when enabled in the cluster configuration.
The priority class values are fixed as per sig-architecture decision and should not be changed:
priorityClasses:
giantswarmCritical:
enabled: true
value: 1000000000
description: "Stuff that the cluster absolutely requires to run and Giant Swarm is adding (e.g., kyverno)"
giantswarmHigh:
enabled: true
value: 900000000
description: "Stuff that should preempt customer workloads (e.g., observability components, crossplane)"Apps should migrate from legacy priority classes:
| Legacy Priority Class | Value | Replace With |
|---|---|---|
crossplane-critical |
600000000 | giantswarm-high |
flux-giantswarm-flux-giantswarm |
1000000000 | giantswarm-critical or giantswarm-high |
prometheus |
500000000 | giantswarm-high |
This chart is designed to work with all Giant Swarm CAPI-based clusters.
- Defined by Giant Swarm sig-architecture
- Reference: giantswarm/roadmap#3483