BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//sf26-agenda//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
UID:SESS-99@sf26-agenda
DTSTAMP:20261103T091709Z
DTSTART:20261104T220000Z
DTEND:20261104T224500Z
SUMMARY:Deep dive into the performance of Queues for Kafka
LOCATION:Room 2002
DESCRIPTION:Apoorv Mittal\, Andrew Schofield\n\nWhen Apache Kafka introduce
 d share groups\, the first question from every architect was: "What's the 
 overhead?" This talk answers that question with data. The talk presents a 
 comprehensive performance analysis of share groups across throughput\, lat
 ency\, and resource utilization under diverse workload profiles. We benchm
 ark share groups against consumer groups in head-to-head comparisons — f
 rom high-throughput streaming workloads to fan-out patterns\, bursty task 
 queues\, and heterogeneous worker pools with variable processing times. Sh
 are groups add record-level state tracking\, acquisition locks\, and per-m
 essage acknowledgment — but how much does that actually cost\, and when 
 does it pay for itself? We show how performance shifts as you scale consum
 ers beyond partition count\, introduce out-of-order acknowledgements\, var
 y message sizes\, and mix fast and slow consumers on the same topic. The r
 esults reveal where share groups match consumer group performance\, where 
 they pull ahead\, and the effect of application design and configuration. 
 You'll leave with a mental model of share group performance characteristic
 s and practical information for designing your applications and tuning you
 r cluster for best performance.
URL:https://sf26.rmoff.net/s/deep-dive-into-the-performance-of-queues-for-k
 afka-99
END:VEVENT
END:VCALENDAR