Kafka can replace a database in some cases, depending on your query patterns and ability to partition the data intelligently. It excels at write-heavy workload

Kafka can replace a database in some cases, depending on your query patterns and ability to partition the data intelligently. It excels at write-heavy workloads and, if you're already using Kafka for streaming or message passing then you can use it as a key-value store in order to avoid the extra operational burden of dealing with a distinct service for caching or storage. But if you need a data warehouse or just a relational query engine for data at rest, or if your data volume is not too large and you're not already using kafka, you're probably better off with a regular DB.