Configure the Batch Engine
Use the Batch Engine when Payload work should be queued and processed in batches instead of running immediately in the current transaction.
This is useful when the Flow, user action, or process can hand off the work and review the resulting Jobs later.
When to use it
Use Batch Engine when you need to:
process a larger number of Payload Jobs asynchronously
control how many Jobs are handled in each batch execution
avoid making the calling Flow wait for each Payload to finish
separate the process that creates work from the process that executes work
monitor queued and completed batch work from one place
For Flow setup, see Use Payloads in Flow.
How Batch Size works
Batch Size controls how many Jobs Payloads should process in each batch execution.
A smaller Batch Size gives each execution less work to do. A larger Batch Size processes more Jobs at once, but each batch execution has more work to complete inside Salesforce limits.
Start with a conservative Batch Size, then adjust after reviewing Job metrics and batch behaviour.
What happens at runtime
When a Payload runs through the Batch Engine, Payloads creates Jobs and marks them as batch work.
Those Jobs stay visible in the Jobs tab. A Job can remain New while it is waiting to be processed. After the Batch Engine processes it, the Job should move to Completed or Failed.
Use the Job record to review the same runtime details you would check for any other Payload run: inputs, request data, response data, target output, error message, and metrics.
Monitor batch work
Open the Batch Engine tab to review current and recent batch activity.
Use it to check whether batch work is waiting, running, or completed. If Jobs stay New longer than expected, check the Batch Engine tab before changing the Payload configuration.
What to check
After testing a batch run, confirm:
the expected Jobs were created
the Jobs are marked as batch work
Batch Size matches the volume and complexity of the Payload work
completed Jobs have the expected request, response, and target output
failed Jobs have clear error messages
metrics do not show unexpected query, DML, CPU, heap, or callout pressure
For Job-level review, see Understand Job records.