After I setup an automation with some heavy Function/Action that perform calls to LLMs, it does hit the limit provided by the LLM providers and generate errors as a consequence.
How can I control the parallelism of calls from Automate ?
In the effect settings you should be able to modify the parallelization settings to help limit rate limits for both logic functions
and actions
.
A personal note is that Batch is “how many object will be my action executed with, as input” and parallelism is “how many actions will be executed in parallel at the same time”.