Skip to main content
Conduktor Gateway advanced pattern Interceptors let you do things standard Kafka can’t: cache repeated fetches, send oversized messages, filter topic content server-side, and collapse many virtual topics onto fewer physical ones. Producers and consumers keep using standard Kafka clients — Gateway handles it at the proxy layer.

​
Available patterns

PatternDescription
Data cachingCache fetch responses at the proxy layer to reduce broker load and repeated consumer reads
Large messagesSplit oversized messages at produce time and reassemble them transparently at consume time
Large batchesBreak produce requests that exceed broker limits into smaller batches automatically
CEL topic filteringDefine logical topics that filter records server-side using Common Expression Language expressions
SQL topicsQuery Kafka topics using SQL semantics — filter, project and transform records at the proxy
Topic concentrationMap multiple virtual topics onto a single physical Kafka topic to reduce partition count and broker overhead

​
When to use advanced patterns

  • Reduce broker load — use caching to serve repeated fetch requests without hitting the broker every time
  • Work around Kafka limits — large message and batch handling let you produce and consume data that exceeds message.max.bytes without reconfiguring brokers
  • Control data exposure — CEL and SQL topic filters let you expose a subset of a topic’s records to specific consumers without duplicating the underlying data
  • Optimize partition usage — topic concentration consolidates logical namespaces onto fewer physical partitions, reducing overhead in multi-tenant environments