getJobs hangs on large Jenkins instances - pagination applied after full enumeration

[sergeykad]

Describe the bug

The getJobs tool hangs indefinitely on Jenkins instances with many jobs, even when using small limit values like 10.

Root Cause

The current implementation in DefaultMcpServer.java materializes ALL jobs before applying pagination:

return parent.getItemsStream()
    .sorted(Comparator.comparing(Item::getName))
    .skip(skip)
    .limit(limit)
    .toList();

The .sorted() operation forces the entire stream to be loaded into memory before skip and limit can be applied. On large Jenkins instances, this causes:

• Excessive memory usage
• CPU spike during sorting
• Timeout before results can be returned

Expected behavior

The getJobs tool should return results within a reasonable time regardless of total job count, since the user only requested a small page of results.

Environment

• Jenkins version: 2.516.1
• MCP Server Plugin version: 0.149.ve5cd2f59de01
• Approximate number of jobs: thousands
• MCP Client: Claude Code CLI

Reproduction steps

1. Have a Jenkins instance with thousands of jobs
2. Call getJobs with limit: 10
3. Observe the call hangs/times out

Suggested fix

Use Jenkins API pagination - Jenkins REST API supports tree=jobs[name]{0,10} style queries that paginate server-side.

Workaround

Users can use parentFullName parameter to scope to a specific folder with fewer jobs, or use getJob with known job names directly.

Related issues

• JENKINS-26746 - "Very slow when there are many folders and projects"
• getJob and searchBuildLog tools return excessive data causing context overflow #147 - getJob returns excessive data (related data volume issue)