Hi, My name is Nandini, a software engineer at Datadog. We run Kafka at large scale with tiered storage enabled, and I've been working on fixes and improvements based on issues we've encountered in production. I recently started contributing to Apache Kafka and wanted to introduce myself and share the JIRAs I've opened:
1. KAFKA-19967 <https://issues.apache.org/jira/browse/KAFKA-19967> / PR <https://github.com/apache/kafka/pull/21102> 2. KAFKA-19968 <https://issues.apache.org/jira/browse/KAFKA-19968> / PR <https://github.com/apache/kafka/pull/21108> 3. KAFKA-19969 <https://issues.apache.org/jira/browse/KAFKA-19969> / PR <https://github.com/apache/kafka/pull/21090> 4. KAFKA-19970 <https://issues.apache.org/jira/browse/KAFKA-19970> / PR <https://github.com/apache/kafka/pull/21089> 5. KAFKA-19971 <https://issues.apache.org/jira/browse/KAFKA-19971> / PR <https://github.com/apache/kafka/pull/21088> These address issues we encountered running tiered storage in production - quota enforcement not being honored correctly, GC pressure from large remote reads, and gaps in observability metrics. I've included unit tests and can share testing results if helpful. I'd appreciate any feedback or reviews. Happy to provide more context, break down changes differently, or jump on a call if that helps. Looking forward to contributing more to Kafka! Thanks, Nandini
