It could be that running a recipe every one minute in a cloud iPaaS is not feasible. What use case you are trying to achieve? An alternate design needs to be in place for this scenario.
What I can perceive is that there are new jobs triggering before the earlier jobs complete their execution. You should always have a design to ensure that you ignore processing of new jobs when any job is executing currently at this time. For this you need to store the execution status of a job in persistent memory and query it to find out if there is any job running. Please reach out to me directly if you need further assistance.
