BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//sf26-agenda//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
UID:SESS-115@sf26-agenda
DTSTAMP:20261020T151709Z
DTSTART:20261105T213000Z
DTEND:20261105T214500Z
SUMMARY:Lifting the Fog: Automated Kafka Topology Discovery for Event-Drive
 n Architecture
LOCATION:Innovation Theater
DESCRIPTION:Swati Bansal\n\nYour Kafka cluster tells you everything about c
 onsumer behavior and nothing about producer identity. That asymmetry is by
  design — but at scale\, it creates a fog that extends every incident\, 
 slows every migration\, and makes orphaned infrastructure impossible to cl
 ean up safely. This session presents a graph-based Application Topology Ag
 ent that reconstructs the complete structural picture from two sources: br
 oker metadata for the consumer side\, and source code configuration for th
 e producer side. Built in one week using AI-assisted development\, the age
 nt identified 42 orphaned consumer groups on first production run and redu
 ced downstream impact triage from hours to 200 milliseconds. This is the a
 rchitecture\, the design decisions\, and the lessons from running it in pr
 oduction.
URL:https://sf26.rmoff.net/s/lifting-the-fog-automated-kafka-topology-disco
 very-for-event-115
END:VEVENT
END:VCALENDAR