Scheduler Failure Handling

InfraLynx treats scheduling failures and job failures differently.

Scheduler Responsibility

• persist schedule definitions
• compute next_run
• enqueue jobs when schedules become due

Job Engine Responsibility

• execute work
• retry failed jobs
• record execution logs
• expose final job state

Missed Run Behavior

In the bootstrap model, due schedules are evaluated on worker polling. If the worker is down, the schedule remains due and is enqueued on the next successful poll after restart.

This favors durability over exact-time guarantees and can lead to delayed execution, but not silent loss.