Pulsar-admin performance degradation

Hi! We have pulsar deployed with Helm on Kubernetes. We upgraded from 2.6.4 to 2.7.2 (and then to 2.9.3) but we found out that the time the brokers take to be ready has increased; we have authentication and authorization enabled (with JWT) and at the first startup the broker generates all the tenants,namespaces and tokens. Starting from 2.7.2 these commands (done through pulsar-admin) have become really slow: the broker takes up to 6 minutes to be ready starting from 2.7.2. Do you have any suggestion? (

Start by reviewing the pulsar-admin commands used during the first broker startup and the broker readiness logs in the Helm-on-Kubernetes deployment. Compare startup behavior across versions 2.6.4, 2.7.2, and 2.9.3, focusing on tenant, namespace, and token generation. Done means identifying the cause of the regression and confirming improved broker readiness time.

Written by the indexing model from the issue text.