---
id: scheduling-tolerations-on
name: "Capability: tolerations settable on ALL workloads (ON)"
description: >-
  Renders the product chart with a toleration configured (key=dedicated, operator=Equal, value=gpu, effect=NoSchedule) and asserts that every workload pod spec carries the toleration. The sample chart does NOT expose a tolerations knob in its values.yaml and has no tolerations block in any Deployment template — setting tolerations has no effect. This scenario is EXPECTED to FAIL (honest red gap) documenting that the chart lacks the tolerations scheduling knob.
labels: {customer: enterprise, profile: compliance}
cluster:
  provider: kind
  k8s_version: v1.30.0
product:
  chart: ./chart
  release: sample
  namespace: sample
  set:
    tolerations[0].key: dedicated
    tolerations[0].operator: Equal
    tolerations[0].value: gpu
    tolerations[0].effect: NoSchedule
asserts:
  - type: scheduling-present
    namespace: sample
    source: rendered
    expect_present: true
    check_nodeSelector: false
    check_tolerations: true
    check_affinity: false
    check_topologySpreadConstraints: false
    tolerations:
      - key: dedicated
        operator: Equal
        value: gpu
        effect: NoSchedule
tags: [capability, compliance]
category: capability
capability: tolerations
tier: capability
mechanisms: ["capability:scheduling"]