[BUG] Content Manager flaky integration tests

[AlexRuiz7]

Describe the bug
The integration tests from the Content Manager often fail on GitHub Runners caused by Java Heap out of memory errors.

These errors are caused by the content manager downloading and indexing the snapshot on each test run.

We need to improve these tests to do not initialize from a snapshot.

Expected behavior
The integration tests set up and mock the environment correctly, skipping or mocking the initialization from snapshots on each test.

Plugins
Content Manager.

Additional context
Example in https://github.com/wazuh/wazuh-indexer-plugins/actions/runs/14858193917/job/41716397598

[AlexRuiz7]

Needs the changes from #401

Initial approach in the attached diff.

patch.txt

[Jorgesnchz]

I've been investigating the issue, to fix it I attempted to mock the initialize method of the SnapshotManager class. However, after reviewing information from OpenSearch, it appears that mocking this method is impossible because it executes within the onNodeStarted method, which runs prior to the integration tests. Therefore, I believe the most viable solution is the implementation of the JobScheduler (#366) which should fix the bug.

[Jorgesnchz]

Testing the bug In a development environment, I was able to reproduce it by following the steps from this comment:

[2025-05-13T14:37:54,232][INFO ][c.w.c.c.CTIClient        ] [integTest-0] Starting snapshot download from [https://cti.wazuh.com/store/contexts/vd_1.0.0/consumers/vd_4.8.0/1931858_1747057658.zip]
»  java.lang.OutOfMemoryError: Java heap space
»  Dumping heap to logs/java_pid146593.hprof ...
»  Heap dump file created [370766792 bytes in 0.719 secs]
» [2025-05-13T12:37:58.243189593Z] [BUILD] Stopping node

I think that we thought It was not happening in a development environment because we forgot to do the steps from the comment previously mentioned.