@Rohit1997jio  Yes it can be fast, but that depends on your dataflow(s). How many other dataflows is your NiFi also running?  All these dataflows share the same JVM resources and utilize server resources.   Also concurrency is important so having a match between number of partitions in your source Topic A and the concurrency within your ConsumeKafkaRecord processor is important.  Concurrent tasks in the ConsumeKafkaRecord you be set using the following formula: (num partitions in Topic A) / (num NiFi nodes in cluster) = (num concurrent tasks set on ConsumeKafkaRecord) Example: Kafka Topic partions = 12 partitions NiFI cluster num nodes = 3 Concurrent tasks set on ConsumeKafkaRecord = 4 4 X 3 = 12 consumers in the consumer group (1  consumer per partition in source topic) Also in NiFI keep in mind that each concurrent task utilizes a thread from NiFi's "Max Timer Driven Thread Count" thread pool. The default pool size is 10.   This means that only 10 concurrent task can execute at the same time.  Generally speaking threads milliseconds, but for optimal performance you'll want to mange your server CPU load average resources and set the pool higher to maximize your throughput performance.   NiFi can show you the Core Load Average (1 = 100% utilization of one physical core).    So if your NiFi server has 4 cores and the load average is 4 or higher, the CPU is saturated.  You'll be able to use this Core Load Average data to determine how much your can increase the size of the "Max timer Driven Thread Count" pool setting.   NiFi recommends setting your "Max Timer Driven Count" initially to 2 - 4 times the number of Physical cores on the NiFi node.  This assume all nodes in the NiFi cluster have same number of physical cores.   If nodes have various numbers of physical cores, use the nodes with the fewest to set your initial pool size.  Then monitor core load average across all NiFi nodes and adjust accordingly.  There is no way to configure different thread pools per node in a NiFi cluster, NiFi expects all all servers in a NiFi cluster to have same hardware configuration.   Please help our community grow. If you found any of the suggestions/solutions provided helped you with solving your issue or answering your question, please take a moment to login and click "Accept as Solution" on one or more of them that helped. Thank you, Matt   ... View more

@Frank168  Glad I was able to identify your issue for you. Can you accept the post that solved your issue.  I see you accepted your response. Thank you, Matt ... View more

@garb  There is an existing Apache NiFi jira reporting this issue here: https://issues.apache.org/jira/browse/NIFI-14729 It aligns with your observations above. Thanks, Matt   ... View more