Substitute dataset-events with the desired topic.
kubectl exec -it kafka-1-0 -- /bin/kafka-run-class kafka.tools.GetOffsetShell --broker-list localhost:9092 --topic dataset-events
You may omit --topic to list all topics, but should then ignore __consumer and _schemas.
kubectl exec -it kafka-1-0 -- /bin/kafka-run-class kafka.tools.GetOffsetShell --broker-list localhost:9092 | grep -v __consumer | grep -v _schemas
dataset-events:0:47472
dataset-events:1:49949
dataset-events:2:49361
dataset-events:3:50122
Substitute topic, partition and offset parameters based on output from command above. Remember that all messages with the same key (often some sort of ID) is put on the same partition.
kubectl exec -it schema-registry-1-0 -- kafka-avro-console-consumer --bootstrap-server kafka-1:9092 --property schema.registry.url=http://localhost:8081 --topic dataset-events --partition 0 --offset 47470
You may redirect the command output to a file and then use jq to get a more readable output.
# Pipe output into file 'kafka-output' while also seeing output in terminal
kubectl exec ... | tee kafka-output
# Show whole messages
cat kafka-output | grep "{" | jq
# Show selected fields in messages
cat kafka-output | grep "{" | jq "{type,fdkId,timestamp}"