Size Kafka topic partitions from throughput and consumer parallelism.
A partition can only be read by one consumer within a group, so consumer count sets a floor on partitions. Adding consumers beyond the partition count leaves them idle rather than increasing throughput. Partitions can be added but never removed from a topic, so under-provisioning is recoverable while over-provisioning is permanent.
Kafka Partition
Partitions = max(throughput ÷ per-partition throughput, consumer count)
Partitions = max(throughput ÷ per-partition throughput, consumer count) A partition can only be read by one consumer within a group, so consumer count sets a floor on partitions. Adding consumers beyond the partition count leaves them idle rather than increasing throughput.
Partitions can be added but never removed from a topic, so under-provisioning is recoverable while over-provisioning is permanent.
This calculator takes 5 inputs: Target throughput, Throughput per partition, Consumers in the group, Retention period, Replication factor. The pre-filled defaults are a realistic starting point — replace them with figures from your own environment for a result you can act on.