I am creating a db for storing sensor data. i have a hypertable for raw readings and continuous aggregate calculating 1 hour aggregations. It is all working well for sensors that are using e.g. average aggregation but now I am lost and don’t know what to do when cumulative sensors (diff aggregation) resets (changed to new or rolled over). In continuous aggregate I will have one bucket that is incorrect (negative or very large value) and I won’t be able fix it by altering row data. Do you have any suggestions that would solve this problem for me?