For example, in Apache IoTDB a metadata template of 90 time series is mounted under root.sg.lu path, and an entity of root.sg.lu.sg(n) is created under root.sg.lu every other day. The entity inserts point data of 90 time series every second. After 100 days, 100 entities will be created. What I want to ask is whether these 100 entities will have 100*90 time series metadata, or will there be only one metadata with 90 time series?
The problem is, according to the above scenario, if there are 100 devices, and each device creates 50 entities per day, there will be about 10036515090=160 million of time series a year. Wonder if this much metadata may affect performance, or is there a command to query the size of the metadata?