ScheduleJobModel

The built-in aggregate that keeps a context's jobs: <context>.schedule, one instance per job in its tenant (the default tenant for a JobScope.SYSTEM job), whose id is the job's name, or, for a JobScope.SHARDED job, one per log partition in the default tenant (shardRef): instanceIds are a job's instances' ids. It belongs to its tenant (no owner), has no lifecycle commands, and decides:

• wow.schedule.Enable(job, rule?): creates the instance or enables it again, with the job's default rule under its own override (the command's, else the one it kept); unchanged, nothing. It schedules its own wow.schedule.Tick under the key tick by that rule, so the firing loop of the partition's owner sends it each slot.

• wow.schedule.Reschedule(rule?): an enabled instance's own override, null to follow the default again.

• wow.schedule.Disable: cancels the schedule.

• wow.schedule.Tick, sent by the schedule: commits wow.Ticked(slot) when the slot is newer than the last one it ran, of the same timer or a later one; an older or repeated slot, or a disabled instance, changes nothing. The tick also reschedules an instance whose rule is no longer its job's default under its override: one that did not override a part follows a change of the default from its next tick. A tick of a job the node no longer has disables the instance.

An Enable that names an unknown job, a job of another scope (a SYSTEM or a SHARDED job outside the default tenant), another instance's id, or an invalid rule, and a Reschedule of an instance that is not enabled, are rejected with FrameworkCodes.SCHEDULE_JOB_INVALID.