Hello,
i was tackling the fifth challenge of the gossip gloomer challenges and I was a little confused. I did pass everything but i dont understand the reasoning behind 2 API endpoints.
We are writing incoming msgs to a log. The messages streams are identifiable by a key String (eg. k1, k2 …). Then a client to the server we are creating can query the logs by polling messages, giving the key and an offset where to start receiving messages. I kept it easy and sent always all previously received messages in one chunk.
There is 2 more API enpoints, CommitOffsets and ListCommitedOffsets. To my understanding these are for Clients to be able to store on the server how far they have read the streams so far. But they are of no consequence to the reading? Also it seems to be okay to store these in a as one atomic counter for all clients and not one CommitedOffset “counter” per client. I thought it would make more sense to have 1 state of CommitedOffset per client, but is this just not tested as it would make the challenge to hard or do i missunderstand the usage of this CommitedOffset?
Thanks for any answer in advance.