Update Helm release descheduler to v0.35.1

[renovate]

This PR contains the following updates:

Package	Update	Change
descheduler	minor	0.34.0 → 0.35.1

---

Release Notes

kubernetes-sigs/descheduler (descheduler)

v0.35.1: Descheduler v0.35.1

Compare Source

What's Changed

• fix(ci): pin helm-unittest plugin version and bump chart-testing-action by @​a7i in #​1834
• [v0.35.0] update helm chart by @​a7i in #​1835
• Automated cherry pick of #​1836: Synchronize helm clusterrole RBAC with base yaml
  #​1826: Add init containers support to Helm chart
  #​1838: Change icon URL in Chart.yaml
  #​1842: fix: resolve detected data races
  #​1847: fix(ci): upgrade codeql-action to v4 and clean up security
  #​1848: update go dependencies
  #​1844: Extend PodLifeTime with condition, exit code, owner kind, by @​a7i in #​1850

Full Changelog: kubernetes-sigs/descheduler@v0.35.0...v0.35.1

v0.35.0: Descheduler v0.35.0

Compare Source

What's Changed

• feat: enable pod protection based on storage classes by @​ricardomaraschini in #​1752
• fix: pod resource calculation to consider native sidecars by @​a7i in #​1771
• docs: fix incorrect gracePeriodSeconds default in README.md by @​petersalas in #​1773
• docs: fix README.md link to kubernetes bot commands by @​Sycrosity in #​1772
• Fix "Current requires cgo or $USER set in environment" error by @​abelfodil in #​1764
• refactor(TestPodLifeTime): remove ineffective owner references assignments by @​ingvagabund in #​1781
• refactor(TestPodLifeTime): have a pod fully created through BuildTestPod without any edits by @​ingvagabund in #​1782
• refactor(TestPodLifeTime): consolidations, simplifications and node instance for each unit test by @​ingvagabund in #​1783
• refactor(TestPodLifeTime): inline pod creation in each unit test to avoid accidental pod spec updates by @​ingvagabund in #​1784
• refactor(TestPodLifeTime): update unit test names and simplify pod creation by @​ingvagabund in #​1785
• feat(TestPodLifeTime): check only expected pods are evicted by @​ingvagabund in #​1787
• feat(PodLifeTime): document the plugin with details that can be used for reasoning during reviews and design discussions by @​ingvagabund in #​1789
• refactor(TestPodLifeTime): split the unit tests into smaller semantically close groups by @​ingvagabund in #​1790
• refactor(TestFindDuplicatePods): have a pod fully created through BuildTestPod without any edits by @​ingvagabund in #​1791
• refactor(TestFindDuplicatePods): reduce duplicates and inline by @​ingvagabund in #​1792
• refactor(TestRemoveDuplicates): reduce test code duplication by @​ingvagabund in #​1793
• refactor(TestRemovePodsHavingTooManyRestarts): inline object creation by @​ingvagabund in #​1794
• refactor(TestPodAntiAffinity): inline object creation by @​ingvagabund in #​1795
• refactor(TestRemovePodsViolatingNodeAffinity): inline object creation by @​ingvagabund in #​1796
• refactor(TestDeletePodsViolatingNodeTaints): inline object creation by @​ingvagabund in #​1797
• doc: introduce contributing guidelines specific to the project by @​ingvagabund in #​1798
• refactor(TestDefaultEvictor): de-dup code and use helpers by @​ingvagabund in #​1803
• refactor(plugins): simplify the way pods are created by @​ingvagabund in #​1804
• fix(TestReadyNodesWithNodeSelector): make sure nodeLister.List always returns a non-empty list so the lister is always tested by @​ingvagabund in #​1800
• refactor(pkg/framework/profile): dedup unit test code by @​ingvagabund in #​1806
• doc(Design Decisions FAQ): Why doesn't the framework provide helpers for registering and retrieving indexers for plugins by @​ingvagabund in #​1807
• feat(profile): inject a plugin instance ID to each built plugin by @​ingvagabund in #​1808
• feat: register a node indexer for the global node selector instead of listing nodes with the selector by @​ingvagabund in #​1802
• chore(pkg/descheduler): make TestPodEvictorReset table driven by @​ingvagabund in #​1810
• refactor(pkg/operator): replace informerResource with a kubeClientSandbox by @​ingvagabund in #​1811
• refactor(pkg/descheduler): more handlers and dropping unused code by @​ingvagabund in #​1813
• refactor(pkg/descheduler): create fake shared informer factory only once by @​ingvagabund in #​1812
• fix(kubeClientSandbox): do not wait for pods in the fake indexers if they are already deleted by @​ingvagabund in #​1814
• test(pkg/descheduler): test a prometheus client update propagates to a plugin profile handle by @​ingvagabund in #​1816
• Add namespace label selector by @​W1seKappa in #​1786
• tests: Prom client testing by @​ingvagabund in #​1818
• Deduplicate descheduler initialization code so unit tests test more of the production code by @​ingvagabund in #​1819
• test(token reconciling): have tests initialize the prom client reconciling through the descheduler's bootstraping entry too by @​ingvagabund in #​1820
• refactor(promClientController): split it into two prom client controllers by @​ingvagabund in #​1821
• feat(pkg/descheduler): create profiles outside the descheduling cycle by @​ingvagabund in #​1815
• refactor: move prometheus client controller related code under a seperate file by @​ingvagabund in #​1823
• Update go dependecies to fix vulnerabilities by @​sammedsingalkar09 in #​1822
• chore: extend the list of supported Go versions by @​ingvagabund in #​1828
• bump(golangci-lint): update and migrate by @​ingvagabund in #​1829
• [v0.35.0] bump to kubernetes 1.35 deps by @​a7i in #​1827
• chore: upgrade github.com/gomarkdown/markdown to latest version by @​a7i in #​1831
• [v0.35.0] update docs and manifests by @​a7i in #​1832
• Change annotations condition to deploymentAnnotations for Deployment object annotations by @​davidandreoletti in #​1830

New Contributors

• @​petersalas made their first contribution in #​1773
• @​Sycrosity made their first contribution in #​1772
• @​abelfodil made their first contribution in #​1764
• @​W1seKappa made their first contribution in #​1786
• @​sammedsingalkar09 made their first contribution in #​1822
• @​davidandreoletti made their first contribution in #​1830

Full Changelog: kubernetes-sigs/descheduler@v0.34.0...v0.35.0

---

Configuration

📅 Schedule: Branch creation - "after 2am and before 8am on monday" in timezone America/New_York, Automerge - At any time (no schedule defined).

🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.

♻ Rebasing: Whenever PR is behind base branch, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about this update again.

---

• If you want to rebase/retry this PR, check this box

---

This PR was generated by Mend Renovate. View the repository job log.

[github-actions]

descheduler (helm) 0.34.0 -> 0.35.1

Risk: 🟢 Safe

The Deep Dive

Update Scope

Helm chart descheduler 0.34.0 → 0.35.1 (includes app image v0.34.0 → v0.35.1). No kustomize image override exists, so the deployed container image updates with this PR. Rendered templates update version labels, image tag, config checksum, and add a new ClusterRole rule for persistentvolumeclaims.

Performance & Stability

• Data race fix (#1842): Resolves detected data races in the descheduler — relevant to all deployments.
• Node indexer optimization (#1802): Uses a node indexer for global node selector instead of listing nodes, improving efficiency in clusters with many nodes.
• Go dependency updates (#1822, #1848): Updated Go dependencies to fix vulnerabilities and bumped to Kubernetes 1.35 client libraries.

Features & UX

• Storage class-based pod protection (#1752): Enables protecting pods based on their storage classes. Not configured — your values.yaml does not use ignorePvcPods or storage class filters. The new PVC RBAC rule in the ClusterRole is added by default but harmless if unused.
• Namespace label selector (#1786): Filter namespaces by labels instead of just names. Not configured — your profiles use namespaces.include with explicit names, not label selectors. Available via namespaceLabelSelector in plugin config if desired.
• PodLifeTime extensions (#1844): PodLifeTime plugin now supports filtering by condition, exit code, and owner kind. Not configured — your PodLifeTime configs use maxPodLifeTimeSeconds and states only. Available if more granular eviction control is desired.
• Init containers support in Helm chart (#1826): Helm chart now supports initContainers in values. Not configured.

Key Fixes

• Native sidecar resource calculation (#1771): Fixes pod resource calculation to properly consider native (init container) sidecars. Relevant if any pods in the cluster use Kubernetes native sidecars — the LowNodeUtilization plugin will now calculate node utilization more accurately.
• Helm ClusterRole RBAC sync (#1836): Synchronizes Helm chart ClusterRole RBAC rules with base YAML manifests, ensuring the Helm-deployed descheduler has all needed permissions.

Hazards & Risks

None identified. The update is a minor version bump with no breaking changes or deprecations. The new ClusterRole rule (persistentvolumeclaims: get/watch/list) is additive and required for the storage class protection feature, but has no impact on existing behavior. All configured plugins (PodLifeTime, RemovePodsHavingTooManyRestarts, LowNodeUtilization, DefaultEvictor) remain compatible with the same configuration format.

Sources

• Descheduler v0.35.0 release
• Descheduler v0.35.1 release
• Storage class pod protection PR
• Data race fix PR
• Native sidecar resource calculation fix

---

🟢 Verdict: Safe

Straightforward minor update with mostly internal refactoring, test improvements, and a data race fix. All configured plugins are fully compatible with no config changes needed. v0.35.1 is the latest release.