[Docs][Core] Add Javadoc to CoordinatorService (scheduling and HA methods)

Sub-issue of #10533

Scope

CoordinatorService is the Zeta master-node coordinator: it schedules pending jobs, handles master failover, and manages JobMaster lifecycle. The job scheduling and HA restoration methods have no Javadoc despite being the most critical for understanding Zeta's reliability model.

File

seatunnel-engine/seatunnel-engine-server/src/main/java/org/apache/seatunnel/engine/server/CoordinatorService.java

Methods to document

Method	Line	What to explain
startPendingJobScheduleThread()	227	Background thread purpose; loop structure; exception handling and restart policy
pendingJobSchedule()	250	Peek-then-consume pattern on PendingJobQueue; resource pre-check before dequeue; PendingSourceState.RESTORE vs SUBMIT semantics
completeFailJob(JobMaster, ...)	334	Why insufficient-resource failures are handled separately from runtime failures
restoreAllRunningJobFromMasterNodeSwitch()	449	Master HA failover: iterates runningJobInfoIMap to reconstruct JobMaster for each in-flight job
restoreJobFromMasterActiveSwitch(Long, JobInfo)	506	Per-job restore path during master switch; difference from normal job submission
checkNewActiveMaster()	539	Hazelcast master-node detection; triggers initCoordinatorService on promotion
clearCoordinatorService()	566	Graceful shutdown ordering: cancels running jobs, shuts down executors, clears IMaps

How to contribute

1. Fork the repo and create a branch
2. Add Javadoc to the methods above
3. Run ./mvnw spotless:apply and ./mvnw -q -DskipTests verify
4. Open a PR with title: [Docs][Core] Add Javadoc to CoordinatorService scheduling and HA methods