Fix endpoint muxer TODOs and make threading use vert.x.

[master312]

#7882

[dmtan90]

🤖 AI Code Review Summary

This is a significant and positive architectural shift from dedicated threads per muxer to a shared Vertx WorkerExecutor pool. This should greatly improve scalability under high load. The move to executeBlocking with a periodic timer is appropriate for this use case. My primary concerns are around native memory management (ensuring all AVPackets in the queue are freed in all circumstances) and the potential CPU overhead of the 10ms periodic timer if the system scales to a large number of concurrent muxers. Please ensure the test suite is robust enough to cover the async nature of the new draining mechanism.

📄 Inline File Comments:

• src/main/java/io/antmedia/muxer/EndpointMuxer.java: The vertx.setPeriodic(10, ...) timer runs every 10ms. This is quite aggressive and could lead to high CPU overhead if many muxers are active. Could this interval be adjusted or made configurable? Also, check if this timer is properly cancelled during all shutdown paths to prevent leaks.
• src/main/java/io/antmedia/muxer/EndpointMuxer.java: LinkedBlockingQueue<AVPacket>: Ensure all paths (especially error cases and teardown) correctly free these AVPacket instances to prevent native memory leaks. The current freeQueuedPackets() helper is a good start, but verify it covers all edge cases.
• src/test/java/io/antmedia/test/MuxerUnitTest.java: The test testTeardownDrainsQueuedPackets (renamed from testShutdownWorkerQueueFullDrainsPacket) likely needs to be carefully verified to ensure it properly mocks/handles the new Vertx-based worker pool lifecycle. Ensure that the test waits long enough for the async drain() task to complete before asserting the queue state.

[burak-58]

I think we may use the available pool. If there is blocking error logs it gives us some information aabout the system.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
89.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud