Optimization for totally unrelated tags


I have a data set that will generate lots of tags (100k+). But the tag will always be filtered and not be used together. For example, a product with 100k+ SKUs. All of the read queries will have a filter statement to get data of only one SKU. No cross-SKU data processing will ever occur.

At first, I think of using a measurement, but I read that measurement behaves like a tag, meaning that they are indexed in the same way.

In terms of using a bucket, I think that using 100k+ buckets seems a little bit odd.

100k+ tag set may be okay for the InfluxDB 2.0 to handle but I think it more or less increases the read query time which I wanted to optimize.

What optimization is suggested in this scenario?