Fix Kubernetes resource name and label length limits for long pipelin…

[Esvanth]

Refactored Naming Logic: Updated K8sPipelineClient.java to ensure all Kubernetes resource names and labels (Jobs, CronJobs, ConfigMaps, Secrets) strictly adhere to the 63-character limit, even with long pipeline names.
Added Unit Tests: Included new test cases in K8sPipelineClientTest.java that verify the truncation logic for pipeline names up to 95 characters.
Successfully Pushed: Your changes are now live on your repository at https://github.com/Esvanth/OpenMetadata.
Next Steps:
If you are the original author of PR 27012, your PR should now show these new commits.
If you are not the author, you can now open a new Pull Request or ask the original author to pull these fixes from your branch: Esvanth/OpenMetadata:fix-27004-pod-name-limit.

[github-actions]

Hi there 👋 Thanks for your contribution!

The OpenMetadata team will review the PR shortly! Once it has been labeled as safe to test, the CI workflows
will start executing and we'll be able to make sure everything is working as expected.

Let us know if you need any help!

[gitar-bot]

🚨 Bug: Label selector mismatch: query uses raw sanitizeName but labels use buildKubernetesName

The buildLabels() method (line 1998) now uses buildKubernetesName("", pipeline.getName()) for the LABEL_PIPELINE value, which truncates long names to 63 chars. However, several methods that query by label selector still use sanitizeName() directly (which no longer truncates):

• killIngestion (line 796): LABEL_PIPELINE + "=" + pipelineName
• getQueuedPipelineStatusInternal (line 898): LABEL_PIPELINE + "=" + pipelineName
• getLastIngestionLogs (line 1083): labelSelectorMap.put(LABEL_PIPELINE, pipelineName)
• deleteJobsForPipeline (called from deletePipeline): LABEL_PIPELINE + "=" + pipelineName

For long pipeline names, the label value stored on the resource is truncated but the label selector used in queries is not, so queries will return zero results. This means: kill won't find running jobs, delete won't clean up jobs, queued status returns empty, and logs can't be retrieved.

Suggested fix:

Use buildKubernetesName("", pipeline.getName()) consistently
for label values and selectors, or extract a shared helper:

private String buildPipelineLabel(String name) {
  return buildKubernetesName("", name);
}

Then use it in both buildLabels() and all query methods.

_{Was this helpful? React with 👍 / 👎 | Reply gitar fix to apply this suggestion}

[gitar-bot]

Code Review 🚫 Blocked 0 resolved / 1 findings

Fixes Kubernetes resource name and label length limits for long pipeline names but introduces a critical label selector mismatch where queries use raw sanitizeName while labels use buildKubernetesName, causing selector failures.

🚨 Bug: Label selector mismatch: query uses raw sanitizeName but labels use buildKubernetesName

📄 openmetadata-service/src/main/java/org/openmetadata/service/clients/pipeline/k8s/K8sPipelineClient.java:796 📄 openmetadata-service/src/main/java/org/openmetadata/service/clients/pipeline/k8s/K8sPipelineClient.java:898 📄 openmetadata-service/src/main/java/org/openmetadata/service/clients/pipeline/k8s/K8sPipelineClient.java:1083 📄 openmetadata-service/src/main/java/org/openmetadata/service/clients/pipeline/k8s/K8sPipelineClient.java:1998

The buildLabels() method (line 1998) now uses buildKubernetesName("", pipeline.getName()) for the LABEL_PIPELINE value, which truncates long names to 63 chars. However, several methods that query by label selector still use sanitizeName() directly (which no longer truncates):

• killIngestion (line 796): LABEL_PIPELINE + "=" + pipelineName
• getQueuedPipelineStatusInternal (line 898): LABEL_PIPELINE + "=" + pipelineName
• getLastIngestionLogs (line 1083): labelSelectorMap.put(LABEL_PIPELINE, pipelineName)
• deleteJobsForPipeline (called from deletePipeline): LABEL_PIPELINE + "=" + pipelineName

For long pipeline names, the label value stored on the resource is truncated but the label selector used in queries is not, so queries will return zero results. This means: kill won't find running jobs, delete won't clean up jobs, queued status returns empty, and logs can't be retrieved.

Suggested fix

Use buildKubernetesName("", pipeline.getName()) consistently
for label values and selectors, or extract a shared helper:

private String buildPipelineLabel(String name) {
  return buildKubernetesName("", name);
}

Then use it in both buildLabels() and all query methods.

🤖 Prompt for agents

Code Review: Fixes Kubernetes resource name and label length limits for long pipeline names but introduces a critical label selector mismatch where queries use raw sanitizeName while labels use buildKubernetesName, causing selector failures.

1. 🚨 Bug: Label selector mismatch: query uses raw sanitizeName but labels use buildKubernetesName
   Files: openmetadata-service/src/main/java/org/openmetadata/service/clients/pipeline/k8s/K8sPipelineClient.java:796, openmetadata-service/src/main/java/org/openmetadata/service/clients/pipeline/k8s/K8sPipelineClient.java:898, openmetadata-service/src/main/java/org/openmetadata/service/clients/pipeline/k8s/K8sPipelineClient.java:1083, openmetadata-service/src/main/java/org/openmetadata/service/clients/pipeline/k8s/K8sPipelineClient.java:1998

   The `buildLabels()` method (line 1998) now uses `buildKubernetesName("", pipeline.getName())` for the `LABEL_PIPELINE` value, which truncates long names to 63 chars. However, several methods that query by label selector still use `sanitizeName()` directly (which no longer truncates):
   
   - `killIngestion` (line 796): `LABEL_PIPELINE + "=" + pipelineName`
   - `getQueuedPipelineStatusInternal` (line 898): `LABEL_PIPELINE + "=" + pipelineName`
   - `getLastIngestionLogs` (line 1083): `labelSelectorMap.put(LABEL_PIPELINE, pipelineName)`
   - `deleteJobsForPipeline` (called from deletePipeline): `LABEL_PIPELINE + "=" + pipelineName`
   
   For long pipeline names, the label value stored on the resource is truncated but the label selector used in queries is not, so queries will return zero results. This means: kill won't find running jobs, delete won't clean up jobs, queued status returns empty, and logs can't be retrieved.

   Suggested fix:
   Use buildKubernetesName("", pipeline.getName()) consistently
   for label values and selectors, or extract a shared helper:
   
   private String buildPipelineLabel(String name) {
     return buildKubernetesName("", name);
   }
   
   Then use it in both buildLabels() and all query methods.

Options

Display: compact → Showing less information.

Comment with these commands to change:

Compact

gitar display:verbose

_{Was this helpful? React with 👍 / 👎 | Gitar}

[Esvanth]

Summary of what was accomplished:
Refactored Naming Logic: Updated K8sPipelineClient.java to ensure all Kubernetes resource names and labels (Jobs, CronJobs, ConfigMaps, Secrets) strictly adhere to the 63-character limit, even with long pipeline names.

Added Unit Tests: Included new test cases in K8sPipelineClientTest.java that verify the truncation logic for pipeline names up to 95 characters.

Successfully Pushed: Your changes are now live on your repository at https://github.com/Esvanth/OpenMetadata.

Next Steps:
If you are the original author of PR 27012, your PR should now show these new commits.
If you are not the author, you can now open a new Pull Request or ask the original author to pull these fixes from your branch: Esvanth/OpenMetadata:fix-27004-pod-name-limit.

#Please do assign the task to me and add an label to it actually safe to go.