Hello InfluxDB Support,
I want to use InfluxDB for the following use-case. Please help me answers questions based on this use-case. I am planning to use InfluxDB on AWS (Probably I can use influx 3.0)
“I have 3000 unique locations and at each location I have 10,000 unique sensors installed. Each sensor is reporting data of Temperature and Pressure. The frequency of reporting of data for each sensor is 5 seconds. My read query pattern is to scan through average & percentile data of both the fields for all the sensors every 30 seconds and load it in-memory.”
Summary
3000 unique locations
10000 unique sensors per location
Writes - 60 million points (3000 locations * 10,000 sensors * 2 fields) per 5 seconds
Reads - Scan through Average & Percentile of fields for all the sensors every 30 seconds.
1.What would be the best schema design for the use-case based on the cardinality and query performance? Also can anyone help with the cardinality of all of this options (is it 60 million different series)?
Following are the options I have thought of. However, I am unsure of how are they in comparison with each other?
a. single bucket , measurement=locationId tags=sensorId fields=Temperature & Pressure
b. single bucket , single measurement (hardcoded) tags=locationId, sensorId fields=Temperature & Pressure
c. single bucket , single measurement (hardcoded) tags=locationId fields=sensorId, Temperature & Pressure
d. Different bucket for different locationId , single measurement (hardcoded) tags=sensorId fields =Temperature & Pressure
e. Suggest something better than this
2.Can Influx DB support such read & write traffic pattern with such high cardinality?
3.Are there any specific configurations in InfluxDB I would have to enable if I would like to support this use-case?
Please let me know if any further information is required from me.
