Millions of measurements vs multiple file/value pairs per measurement


Hi, starting with influxdb for our massive data project. We will be getting huge amount of of different measurements from different clients. Will be writing search queries to show all these measurements for different clients and showing visual graphs per measurement. Question: is it better to design the schema with millions measurements/tables(example: timestamp, valueOfMeasurement) OR one measurement/table but multiple field/values.( example: timestamp, measurement1=10. measurement=30, measurement3=100…). What are best practices? Appreciate your help!