I have two high-level questions broken up into more individual questions, both of the high-level questions deal with a consumer group an Apache Kafka Streams API is creating and using.
First of all, is the output from the kafka-consumer-group.sh script. I am getting weird output that doesn't really tell me where a particular consumer is at though they seem to be connected to a particular group/topic/partition:
TOPIC PARTITION CURRENT-OFFSET LOG-END-OFFSET LAG
STANDARD_DATA 9 11 11 0 myConsumer-7fc71848-465b-4817-93b3-42b9ba290dcd-StreamThread-1-consumer-4fd9dc15-d8a7-4598-85a9-3761ae6a747b/1.1.1.1 myConsumer-7fc71848-465b-4817-93b3-42b9ba290dcd-StreamThread-1-consumer
STANDARD_DATA 0 4 11 7 myConsumer-13b61e5a-6289-45db-844b-3ef8c5a26782-StreamThread-5-consumer-28e1c7bf-860d-44d6-bf58-5e0ff875587c/1.1.1.1 myConsumer-13b61e5a-6289-45db-844b-3ef8c5a26782-StreamThread-5-consumer
STANDARD_DATA 4 - 10 - myConsumer-7fc71848-465b-4817-93b3-42b9ba290dcd-StreamThread-4-consumer-a3023af6-eafb-4633-85f1-048c20c4dfb3/1.1.1.1 myConsumer-7fc71848-465b-4817-93b3-42b9ba290dcd-StreamThread-4-consumer
STANDARD_DATA 5 - 10 - myConsumer-7fc71848-465b-4817-93b3-42b9ba290dcd-StreamThread-3-consumer-a81f1399-1fc4-4579-b24f-fa8fee01fabf/1.1.1.1 myConsumer-7fc71848-465b-4817-93b3-42b9ba290dcd-StreamThread-3-consumer
STANDARD_DATA 3 - 12 - myConsumer-13b61e5a-6289-45db-844b-3ef8c5a26782-StreamThread-2-consumer-6a83bfcc-2c6e-4e9d-a819-029ac8c6ae17/1.1.1.1 myConsumer-13b61e5a-6289-45db-844b-3ef8c5a26782-StreamThread-2-consumer
STANDARD_DATA 8 12 12 0 myConsumer-13b61e5a-6289-45db-844b-3ef8c5a26782-StreamThread-4-consumer-6d46bed3-70c4-4c7f-8e53-f9591192bc3f/1.1.1.1 myConsumer-13b61e5a-6289-45db-844b-3ef8c5a26782-StreamThread-4-consumer
STANDARD_DATA 7 - 11 - myConsumer-13b61e5a-6289-45db-844b-3ef8c5a26782-StreamThread-3-consumer-5313315b-ded9-4fe7-ac9d-d8d5b20dd5b9/1.1.1.1 myConsumer-13b61e5a-6289-45db-844b-3ef8c5a26782-StreamThread-3-consumer
STANDARD_DATA 2 10 10 0 myConsumer-b9402faf-4b37-479f-82be-a17eaa180c62-StreamThread-1-consumer-c08a648f-548e-47a8-8bc5-7b6fa3bc1fb5/1.1.1.1 myConsumer-b9402faf-4b37-479f-82be-a17eaa180c62-StreamThread-1-consumer
STANDARD_DATA 1 2 10 8 myConsumer-7fc71848-465b-4817-93b3-42b9ba290dcd-StreamThread-2-consumer-08d99679-d430-4e9f-a3b9-11e558ca34a4/1.1.1.1 myConsumer-7fc71848-465b-4817-93b3-42b9ba290dcd-StreamThread-2-consumer
STANDARD_DATA 6 - 12 - myConsumer-7fc71848-465b-4817-93b3-42b9ba290dcd-StreamThread-5-consumer-666040f8-d4d0-49e9-9db6-c6efee49ebe1/1.1.1.1 myConsumer-7fc71848-465b-4817-93b3-42b9ba290dcd-StreamThread-5-consumer
- Why is it that some CURRENT-OFFSETS (3rd column) and LAG (4th column) are showing up as '-' when I can query the API of Kafka directly to distinguish that they are in fact caught up?
(queried through golang API)
4 myConsumer-7fc71848-465b-4817-93b3-42b9ba290dcd-StreamThread-4-consumer-a3023af6-eafb-4633-85f1-048c20c4dfb3 OFFSET: 10 LOG-END: 10 LAG: 0
- Also, why on would that offset not being showing up as represented in the logs (aka, it should be caught up) in general?
My second high level question is that of streams. We have a stream process working that is, at random times (mostly during a restart), resetting to the earliest offset available in particular topics. Throughout the code there is no 'reset', and the OFFSET_RESET isn't touched. I also can confirm that we are not using the 'exactly-once', so I am not sure where exactly these offset resets are coming into play.
Once again, its basically:
stream process is churning through the data, something ~happens~ and then our offsets are back to ground 0, processing again. This can be going on for days to weeks before it decides to reset too, so the committing of offsets IS happening.