Hello Druids. Our team is considering an upgrade from version 27.0.0 to 30.0.0. When running ingestions we're seeing ingestions in 30.0.0 being slower for some datasources by about 10%. For context I am comparing by running ingestion of the s3 paths in a 30.0.0 cluster and our current 27.0.0 cluster. For the 30.0.0, we are creating a brand new Data source where as in 27.0.0 the ingestion is happening on a pre existing data source. Whats strange is that for some ingestions it is also running faster. Are there any changes in 30.0.0 which can cause ingestions for some datasources be slower? These are all batch ingestions.

Is this Hadoop based, JSON based or MSQ?

MSQ

Nope. not aware of any slowness. maybe you enabled durable storage?

Hey @Abhishek Agarwal, I think this would be some kind of connection issue as the times seem to have stabilised now. Seems like some one off.