Do you know if by workflow, there is a limitation about the number of “Kafka Trigger” ?
My issue is, i defined 4 “Kafka Trigger” in one workflow using 4 different topic, three of them can received data and for the last one seems to have been removed from his broker. The group id is based on the workflow id, if i check the lag value for this particular group id it’s not set to 0, which means there is several message to read.
I don’t know if it’s relevant but I have some issue about kafka in my server logs like : {"level":"ERROR","timestamp":"2025-06-26T05:32:22.382Z","logger":"kafkajs","message":"[Connection] Connection timeout","broker":"yyy.confluent.cloud:9092","clientId":""}
clientId is “” : difficult to debug
If i create a new workflow with a “Kafka Trigger” using the same topic, it works ! The message are processed by the workflow.
Is there a reason you want all of your consumers to share the same consumer-group-id? Having multiple triggers (which are single-topic by design) all in the same consumer-group might be part of the reason things aren’t working as expected. Kafka tries to balance/track all connected consumers in a consumer-group to distribute messages between them. Using multiple triggers with the same consumer-group-id, on different topics, might not have the desired effect on the number of consumers n8n connects. It’s also not very well documented how n8n does or doesn’t combine subscriptions to multiple topics in each consumer process. I have found it to be more predictable to have as little overlap as possible between consumer-groups, topics, and consumers. It might help to review how Kafka uses consumer-groups.
If your concern is duplicating the workflow logic, you could “do what works” and put each trigger in its own 2-step workflow which just receives events/messages and passes them to a common workflow (with the processing logic) using an Execute workflow node.
