Regarding that implementation, would you say it is better to split the workflow in smaller bits ? End the first workflow when the file is submitted for processing and another one when I am signaled the end of it ? Would it make any difference in regards to temporal ?
See this post.