Discussion: Strategy to handle data skew in RisingWave

[kwannoel]

1. Strategy to diagnose data skew in user system. Should be a quick tool (e.g. some SQL) we can use to easily check data skew.
2. Strategies to fix data skews.

We should test and document these, and let users also check their own cluster for these cases.

[BugenZhao]

1. Should be a quick tool (e.g. some SQL) we can use to easily check data skew.

I suppose this can be achieved by comparing some metrics (like throughput or blocking rate) between the actors in the same fragment. We could visualize it in the dashboard UI.

[github-actions]

This issue has been open for 60 days with no activity. Could you please update the status? Feel free to continue discussion or close as not planned.