| Provider | Category |
|---|---|
| Confluent | SaaS, Platform |
The number of partitions configured for a topic can impact the ingress and egress rates, producer and consumer performance. In general, more partitions will lead to better performance. However, a very large number of partitions can impact the performance of Kafka brokers, and too many partitions may lead to poor performance or instability.
This policy can be used to limit the number of partitions that may be assigned
to topics, both from a minimum and maximum level. These are configured using
the topic_partitions_min and topic_partitions_max parameters, and all topics
must include partition counts within that range.
There may be circumstances where differing partition ranges may be suitable,
for example, test topics (topics begining with a test. prefix) may have a
partition range of 1 to 3 partitions, while production topics (topics
begining with production. prefix) have a partition range of 3 to 12
partitions. If this is the case, a custom policy could be used.