Standalone
When a job only needs to run on a single machine, the standalone execution model keeps things simple: SchedulerX dispatches each job instance to one randomly selected worker. No sharding logic, no broadcast fan-out -- just one worker runs the entire job from start to finish. Standalone supports all job types in SchedulerX.
How it works
Each time a job triggers, SchedulerX performs the following sequence:
Select a worker -- SchedulerX randomly picks one available worker from the connected pool.
Dispatch the job instance -- The job instance is sent to the selected worker.
Run to completion -- The selected worker runs the job instance in full. No work is distributed to other workers.
Successive runs of the same job may land on different workers because worker selection is random per trigger.
When to use standalone
Standalone is the simplest execution model. Use it when the entire workload fits on a single worker.
| Scenario | Example |
|---|---|
| Periodic maintenance | Data cleanup, cache refresh, report generation |
| Lightweight processing | Jobs where one worker has enough capacity |
| Single-purpose tasks | Any job that does not need to fan out across workers |
If your workload requires running on every worker simultaneously, consider the broadcast model. If your workload is too large for one worker and needs to be split, consider the sharding model.