How Job queues are processed in Ensemble

Hi Guys,

I have a scheduled task that run every 5 mins to execute a function with 10 things to do, issue that those 10 thing are processed in a loop that takes more than 7 mins to finish, so basically I would like to understand what will happen when the task will run in next 5 mins given that the 10 things to do in my function from the first run didn't finish yet, would it ignore what's left from the 10 things to do and start again or the first run or would the first run  continue to execute in some job # and the new run will run in parallel with new Job#, and if that's the case would this affrect the system performance or is it better to expand my scheduled task eg. to 10 mins to ensure that all 10 things has been completed ?

Thanks