fix(scheduler): qualify SQL retry templates to fix multi-bean startup failure

[bishnubista]

Summary

When the scheduler Postgres or MySQL persistence module is loaded alongside the full Conductor server context, startup can fail because multiple RetryTemplate beans are present: core's onTransientErrorRetryTemplate and the DB-specific postgresRetryTemplate / mysqlRetryTemplate. The scheduler SchedulerDAO and SchedulerArchivalDAO constructors autowired RetryTemplate by type, which is ambiguous and can cause a NoUniqueBeanDefinitionException during context refresh.

This change qualifies the constructor parameters in both scheduler persistence modules so the correct DB-specific retry template is injected:

• PostgresSchedulerConfiguration: schedulerDAO and schedulerArchivalDAO now require @Qualifier("postgresRetryTemplate").
• MySQLSchedulerConfiguration: schedulerDAO and schedulerArchivalDAO now require @Qualifier("mysqlRetryTemplate").

The shared smoke-test fixture now reproduces the multi-bean context by registering core's onTransientErrorRetryTemplate plus each DB-specific retry template. The tests assert both SchedulerDAO and SchedulerArchivalDAO are created with the expected concrete classes and verify both DAOs hold the DB-specific RetryTemplate instance.

Testing

Fail-first verified before applying the production fix: both smoke tests failed with two RetryTemplate beans (onTransientErrorRetryTemplate plus postgresRetryTemplate / mysqlRetryTemplate). After the fix:

./gradlew :conductor-scheduler-postgres-persistence:test --tests '*PostgresSchedulerAutoConfigurationSmokeTest'
./gradlew :conductor-scheduler-mysql-persistence:test --tests '*MySQLSchedulerAutoConfigurationSmokeTest'
./gradlew :conductor-scheduler-postgres-persistence:test --tests '*PostgresSchedulerAutoConfigurationSmokeTest' \
          :conductor-scheduler-mysql-persistence:test --tests '*MySQLSchedulerAutoConfigurationSmokeTest'
./gradlew spotlessApply
git diff --check

Docker Compose validation with Postgres:

YARN_OPTS='--network-timeout 10000000' docker compose -f docker/docker-compose-postgres.yaml up -d --build

• Docker build completed successfully.
• conductor-server and conductor-postgres reached healthy state.
• curl http://localhost:8000/health returned HTTP 200 with healthy=true.
• Scheduler persistence auto-configuration was exercised: Postgres contained the scheduler, scheduler_execution, and scheduler_next_run tables, and flyway_schema_history_scheduler showed successful rows for the baseline plus migrations 1 and 2.
• Note: /api/scheduler/schedules returned 404 in the vanilla compose setup, so Docker verification was scoped to the scheduler persistence auto-configuration path via migrations and tables rather than scheduler HTTP calls.
• Teardown: docker compose -f docker/docker-compose-postgres.yaml down -v.

Fixes

Fixes #1084

[bishnubista]

@v1r3n Can you please review?