Closed rtyler closed 9 years ago
So I can answer this now, here's the answer for posterity!
On JRuby, since we're relying on the "high level consumer" API from Kafka's client libraries, we cannot get access to the offset being consumed. It is not exposed. it's possible to get the offset when implementing a "simple" (read: basic) consumer but then the user is responsible for implementing their own leader following and error handling code. I'm exploring this work in my beetle experimenting, which might prove useful at some point in the future here.
For MRI, I believe we can get at the offsets, but I do not believe it is worth introducing without API symetery between JRuby and MRI
Thinking aloud, I can imagine certain usecases where the consumer might want to hold onto their current offset, partition tuple for logging, etc.