gavinchou commented on issue #64700: URL: https://github.com/apache/doris/issues/64700#issuecomment-4768644020
The key evidence is still missing here. To determine whether this is a Doris 4.1.2 regression or an environment/workload/client-side effect, we need a controlled comparison using the same payload size, same target table, and the same Stream Load parameters before and after the upgrade. Please provide the full Stream Load JSON responses for several slow labels, especially the timing fields such as `LoadTimeMs`, `BeginTxnTimeMs`, `StreamLoadPutTimeMs`, `ReceiveDataTimeMs`, `ReadDataTimeMs`, `WriteDataTimeMs`, and `CommitAndPublishTimeMs`. Matching BE logs for `finished to execute stream load` and FE logs around begin/commit/publish for those labels would also be very helpful. From the pasted application logs alone, the after-upgrade batches are much larger than the before-upgrade batches, so the current evidence does not by itself prove a 10x Doris write regression. It does show higher observed latency, especially for small Stream Load requests, but the server-side timing fields are needed to locate where the extra time is spent. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
