The number of rows of UDF input chunk is too small

[StrikeW]

Describe the bug

https://grafana.prod.risingwave.cloud/d/EpkBw5W4k/risingwave-dev-dashboard?from=1705143551000&orgId=1&to=1705159885000&var-component=All&var-datasource=P7AEF1E9BD6C1839A&var-instance=risingwave&var-namespace=rwc-g1hhgcv4t0f0k8cr19qbl8coqk-test&var-pod=All&var-table=All

Error message/log

No response

To Reproduce

No response

Expected behavior

No response

How did you deploy RisingWave?

No response

The version of RisingWave

v1.6-da175e8

Additional context

https://risingwave-labs.slack.com/archives/C0606NNR74P/p1705155772714269?thread_ts=1705135558.026809&cid=C0606NNR74P

[StrikeW]

cc @TennyZhuang @wangrunji0408

[BugenZhao]

I suppose this issue is not exclusive to UDF. It actually depends on the chunk size from the upstream executor. For instance, if there's an aggregation with few groups in the upstream, then small chunks for evaluation is inevitable.

May I ask about the throughput of the UDF?

[StrikeW]

You can check the grafana link to see the udf throughput and other metrics.

[wangrunji0408]

Is it worthy to merge small chunks into bigger ones at the entry of ProjectionExecutor? or do it in a new executor that can be connected before the projection?

[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.