Hi
after an error, a fail or anything that may interrupt the executor, the stats calculated at the end of the pipeline doesn't takes in account the stats of already accomplished job
I think it will be better if the final stats are the aggregated stats of all the tasks not only the one that just run this time
Thanks for your work, it's a pleasure to work with this tool
Best Regards