For some context, I am using Kafka’s docker-compose configuration file to host my Kafka cluster.
Seperately, 2 containerized python clients, each with the confluent-librdkafka-plugins installed, 1 producer and 1 consumer.
If I understood correctly, to have access to the monitoring interceptor data, I just added the following key-value to their configuration files so the data is inserted into the _confluent-monitoring topic: ‘plugin.library.paths’: ‘monitoring-interceptor’
I checked and the topic has definitely been created on my brokers, but I don’t seem to have the end-to-end latency information I wanted on my control-center UI.
Am I missing something in my setup?
Thank you in advance for any information
@landaudiogo have you seent his section of the docs on “librdkafka-based” monitoring interceptors?
https://docs.confluent.io/platform/current/control-center/installation/clients.html#librdkafka-based-clients
The process for Python (which is librdkafka based), is different than for Java applications.
Although the documentation doesn’t really provide any detailed information regarding python clients when containerized, I think I did manage to install the confluent-kafka-plugins correctly in each container.
Now I am just not sure whether I have to do anything else to the clients configurations other than adding the “plugin.library.paths”: “monitoring-interceptor”
The reason why I believe the configuration might be working, was that before I had the monitoring-interceptor, in the consumption tab in the control-center, there was a link indicating that to have that information available I had to install the interceptors, which no longer appears… But the problem is that the data is coming back as null to the control-center, providing with empty graphs
@landaudiogo you could try setting the following configuration for more debugging information:
'confluent.monitoring.interceptor.icdebug': True