[BanyanDB-Server] Fix TopN High Cardinality

[hanahmily]

Search before asking

• I had searched in the issues and found no similar feature requirement.

Description

I have analyzed the series index data distribution in the skywalking official demo. I found that the TopN related result measures suffer high cardinality issues.

The default candidate size of each TopN list is 1000, which takes up too much series index space. I propose to move rankNum (0,1,2,...1000) from the entity items to the nanoseconds position in the timestamp. The timestamps are down-sampled based on the interval defined in the measure's schema.

data-0-sidx.csv

Use case

No response

Related issues

No response

Are you willing to submit a pull request to implement this on your own?

• Yes I am willing to submit a pull request on my own!

Code of Conduct

• I agree to follow this project's Code of Conduct