After a bad deploy, analysts need to re-read last Tuesday's sensor topic. What Kafka attribute makes that possible within the configured window?

Re-reading last Tuesday’s sensor topic after a bad deploy is possible because of durable log retention: events remain on the topic for a configured retention period and can be replayed by resetting offsets. Immediate permanent deletion after one consumer sees a message would prevent recovery. Volatile memory-only topics with no disk persistence lose history on restart. Guaranteed infinite storage of every byte forever on the broker alone is not how retention windows work and confuses Kafka with a data lake. Exam caveat: retention is time/size bounded—seven years of raw telemetry does not belong solely on broker disks. Operational check: within retention, reset a consumer group to last Tuesday’s offsets, reprocess the sensor topic, and confirm rebuilt outputs match after the fix. In production terms for Kafka retention for tactical replay, assign an owning team, wire a dashboard that shows the pattern is healthy under peak civic load, and require a staging drill that rehearses failure and recovery before resident impact. Revisit the choice when traffic shape, retention, or integrity needs change, and keep a named rollback owner on the change record. Re-test both the happy path and the failure path for Kafka retention for tactical replay after material dependency upgrades so regressions are caught before the next public release.