Every time message or record is assigned to a partition in Kafka, it is provided with an offset.
The offset denotes the position of the record in that partition.
A record can be uniquely identified within a partition using the offset value.
The partition offset only carries meaning within that particular partition. Records are always added to the ends of partitions, and therefore, older records will have a lower offset.
https://www.learningjournal.guru/courses/kafka/kafka-foundation-training/offset-management/