feat(processing_engine): Store triggers by DbId and TriggerId instead of name

[kumarlokesh]

Attempts to close ##26153

Describe your proposed changes here.

1. Fixed type safety in PluginChannels struct:

   • Changed schedule_triggers type from HashMap<String, HashMap<String, mpsc::Sender<ScheduleEvent>>> to HashMap<(DbId, TriggerId), mpsc::Sender<ScheduleEvent>> to properly handle schedule events

2. Improved trigger removal logic in remove_trigger method:

   • Removed incorrect attempt to get and remove from trigger maps
   • Now directly removing entries from HashMaps using correct keys
   • For WAL and schedule triggers: using tuple key (db_id, trigger_id)
   • For request triggers: using path string directly

• I've read the contributing section of the project README.
• Signed CLA (if not already signed).

[kumarlokesh]

It seems the test cli::test_trigger_cache_cleanup is failing https://app.circleci.com/pipelines/github/influxdata/influxdb/44923/workflows/015cf4c0-526b-44db-8bfa-e5c806f56648/jobs/425503:

> ──── STDERR:             influxdb3::lib cli::test_trigger_cache_cleanup
> 
> thread 'cli::test_trigger_cache_cleanup' panicked at influxdb3/tests/cli/mod.rs:2548:5:
> assertion failed: `(left == right)`
> 
> Diff < left / right > :
>  Array [
>      Object {
> <        "count(*)": Number(2),
> >        "count(*)": Number(1),
>      },
>  ]

Need to investigate this further.