When the "-e" or "--ends_at_latest_offsets" is added as an argument the application terminates when the end of the topic is reached.
The end of the topic is determined by storing the highest offsets known on startup. When a partitions reaches this offset, the partition is unassigned from the consumer.
What happens if a rebalance triggers while certain partitions are already unassigned? The rdkafka docs say that auto rebalance is disabled when you use assign, but not on unassign
How will multiple threads work? To by knowledge, kafka consumers are single threaded. I did not find any initialization of worker threads in my time looking at the codebase, although threads are mentioned in some places.
What happens when reading from Azure Event Hubs, does the (un)assigning of partitions work the same?
When the "-e" or "--ends_at_latest_offsets" is added as an argument the application terminates when the end of the topic is reached. The end of the topic is determined by storing the highest offsets known on startup. When a partitions reaches this offset, the partition is unassigned from the consumer.
Fixes #75
Curious to hear any feedback!