[KAFKA-16548] Avoid decompressing/collecting all records when all we want to do is to find a single matched record from remote storage

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 3.8.0
• Assignee: 黃竣陽
• Reporter: Chia-Ping Tsai
• Created: 15/Apr/24 08:21
• Updated: 22/Apr/24 13:23
• Resolved: 22/Apr/24 13:23

Description

https://github.com/apache/kafka/blob/437ebb941e1485a6f1ad7afbba3814bfa310c9cf/core/src/main/java/kafka/log/remote/RemoteLogManager.java#L465

`DefaultRecordBatch#iterator` will decompress all records to a new collection when those records are in compression. That is a extra and unnecessary cost when we do early return.

In short, we should use `streamingIterator` rather than `iterator` in this case.

Issue Links

• links to: GitHub Pull Request #15765