Jo Wijnant
09/24/2025, 1:38 PMproperties.client.id on a source table (kafka consumer) is ignored whereas on a sink table (kafka producer) it effectively applies.
This is the relevant SQL for all my tables:
CREATE TABLE ...
WITH (
'connector' = 'upsert-kafka',
'properties.client.id' = 'flink-project-usage',
...)
The client.id is applied to the producer (logs of taskmanager):
2025-09-24 13:20:17,676 INFO org.apache.kafka.common.config.AbstractConfig [] - ProducerConfig values:
acks = -1
...
client.id = flink-project-usage
But not for the consumer. It takes {group.id}-{index}
2025-09-24 13:20:17,478 INFO org.apache.kafka.common.config.AbstractConfig [] - ConsumerConfig values:
allow.auto.create.topics = false
...
client.id = coz_irods_project_usage-0
Why can I not set client.id on consumer?
I like to label all consumers with the same client.id so I can easily define kafka quotas for this client. But now I'd have to define multiple quotas (coz_irods_project_usage-0, coz_irods_project_usage-1, coz_irods_project_usage-2 ...)