[JENKINS-76265] Memory leak on not closed clients

[jenkins-infra-bot]

Using latest (4384) version of the Kubernetes plugin we observed memory leak on the Jenkins. Analyzing heap dump found that it is caused by Kubernetes clients that were marked as expired but were never closed.

From comments of other people it seems to be cause by this change:

#1747

 

logs from KubernetesClientProvider: (Closing expired client log is never present)

Nov 04, 2025 1:03:31 PM FINE org.csanchez.jenkins.plugins.kubernetes.KubernetesClientProvider
Expiring Kubernetes client mycluster io.fabric8.kubernetes.client.impl.KubernetesClientImpl@​2d6bac84: EXPIRED
Nov 04, 2025 1:03:31 PM FINE org.csanchez.jenkins.plugins.kubernetes.KubernetesClientProvider
Created new Kubernetes client: mycluster io.fabric8.kubernetes.client.impl.KubernetesClientImpl@​77b7e74d
Nov 04, 2025 1:15:42 PM FINE org.csanchez.jenkins.plugins.kubernetes.KubernetesClientProvider
Expiring Kubernetes client mycluster io.fabric8.kubernetes.client.impl.KubernetesClientImpl@​77b7e74d: EXPIRED
Nov 04, 2025 1:15:43 PM FINE org.csanchez.jenkins.plugins.kubernetes.KubernetesClientProvider
Created new Kubernetes client: mycluster io.fabric8.kubernetes.client.impl.KubernetesClientImpl@​40538e0c
Nov 04, 2025 1:41:12 PM FINE org.csanchez.jenkins.plugins.kubernetes.KubernetesClientProvider
Expiring Kubernetes client mycluster io.fabric8.kubernetes.client.impl.KubernetesClientImpl@​40538e0c: EXPIRED
Nov 04, 2025 1:41:13 PM FINE org.csanchez.jenkins.plugins.kubernetes.KubernetesClientProvider
Created new Kubernetes client: mycluster io.fabric8.kubernetes.client.impl.KubernetesClientImpl@​29bdc093

---

Originally reported by krzysztof_zaremba, imported from: Memory leak on not closed clients

• status: Open
• priority: Critical
• component(s): kubernetes-plugin
• resolution: Unresolved
• votes: 0
• watchers: 2
• imported: 2025-12-02

Raw content of original issue

Using latest (4384) version of the Kubernetes plugin we observed memory leak on the Jenkins. Analyzing heap dump found that it is caused by Kubernetes clients that were marked as expired but were never closed.

From comments of other people it seems to be cause by this change:

#1747

 

logs from KubernetesClientProvider: (Closing expired client log is never present)

Nov 04, 2025 1:03:31 PM FINE org.csanchez.jenkins.plugins.kubernetes.KubernetesClientProvider
Expiring Kubernetes client mycluster io.fabric8.kubernetes.client.impl.KubernetesClientImpl@2d6bac84: EXPIRED
Nov 04, 2025 1:03:31 PM FINE org.csanchez.jenkins.plugins.kubernetes.KubernetesClientProvider
Created new Kubernetes client: mycluster io.fabric8.kubernetes.client.impl.KubernetesClientImpl@77b7e74d
Nov 04, 2025 1:15:42 PM FINE org.csanchez.jenkins.plugins.kubernetes.KubernetesClientProvider
Expiring Kubernetes client mycluster io.fabric8.kubernetes.client.impl.KubernetesClientImpl@77b7e74d: EXPIRED
Nov 04, 2025 1:15:43 PM FINE org.csanchez.jenkins.plugins.kubernetes.KubernetesClientProvider
Created new Kubernetes client: mycluster io.fabric8.kubernetes.client.impl.KubernetesClientImpl@40538e0c
Nov 04, 2025 1:41:12 PM FINE org.csanchez.jenkins.plugins.kubernetes.KubernetesClientProvider
Expiring Kubernetes client mycluster io.fabric8.kubernetes.client.impl.KubernetesClientImpl@40538e0c: EXPIRED
Nov 04, 2025 1:41:13 PM FINE org.csanchez.jenkins.plugins.kubernetes.KubernetesClientProvider
Created new Kubernetes client: mycluster io.fabric8.kubernetes.client.impl.KubernetesClientImpl@29bdc093

[jenkins-infra-bot]

markewaite:

• Original comment link
• Raw content of original comment:

  Please provide enough details so that others can duplicate the issue.  "How to report an issue" provides a checklist of the types of information that need to be included in the issue report.
  

Please provide enough details so that others can duplicate the issue. "How to report an issue" provides a checklist of the types of information that need to be included in the issue report.

[hottea773]

We hit the same thing; eventually caused our CPU to skyrocket, presumed to be due to GC. Downgrading to 4358 has fixed it.