majialoong commented on PR #21291: URL: https://github.com/apache/kafka/pull/21291#issuecomment-3745543247
> > Could we add a slight warning to the user regarding using high value of updates/snapshot as that could stall cleanup for high traffic partitions. Similarly low value for a share partition with huge amount of state (state batches) which would cause repeated share snapshots with large amount of data. > > I've had a change of heart. If weird config values can blow up a user's cluster, we should lock that door tight. Prevention is better than a cure. I suggest increasing the lower bound now to avoid the need for a future KIP (similar to KIP-1030). For example, could we use `between(200, 1000)` instead? @majialoong @smjn WDYT? As @smjn mentioned, this parameter can cause issues when configured too high or too low. However, without knowing users' specific scenarios, it's difficult to define a universally reasonable range. If we set an upper limit of 1000: for high-traffic scenarios, users may still experience frequent snapshot generation even when configured at the maximum value, but would be unable to further increase it to optimize performance. They would have to modify the code to bypass the limit, which increases the operational burden. -- 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]
