This message was deleted.

Thank you for sharing the article for better understanding @Yu Wei Sung. As from what I understood, for the bookkeeper configuration set by us, we should expect throughput on at least 5 bookies out of 15. But we can see in the graph that traffic is coming on only 2 bookies. Same configuration example is covered in shared article. Please correct me if I misunderstood something.

two topics, Qw=3, Qa=1, if the ledger is not closed, this is expected

The question becomes how frequently the ledger is rolling over. It should happen rather frequently at a rate of 1M records a second. Can you share the internal topic stats? Focusing on the number of ledgers in the topic. Also, what are the topic’s data retention and/or offloading policies?

Hi @Sandesh Shingare. Any input on @David K’s message above? We are also happy to set up a call to discuss your challenges if that would help.

Hello @Yu Wei Sung @David K @Silvia Santillan Nava, We have figured out the issue in our pulsar test setup. Due to a manual error, client application was deployed with default namespace having "bookkeeperEnsemble" : 1, "bookkeeperWriteQuorum" : 1, "bookkeeperAckQuorum" : 1. Data was being written to different namespace altogether which caused the confusion. Apologies for the confusion. And once again thank you for your prompt support. 🙌 Will get in touch again in case of any other doubt or issue. ✌️