pulsar

<@U9NHC1ETG> thank you for replying. we are using apachepulsar/pulsar-all:2.10.2 image for the broker. we installed the cluster using the master branch of pulsar helm chart repository last week.

The command that we ran from open messaging benchmark tool is below
```sudo bin/benchmark   --drivers driver-pulsar/mypulsar.yaml   workloads/10k-topic-1kb-4p-4c-100k.yaml```

There is a known performance limitation in the metadata store when you exceed ~10K topics in the same namespace (which I think is what you are hitting). There is a fix for this in Pulsar 2.12

good to know. will try with 2.12 tomorrow and let you know. Thanks a lot!

seems like 2.12 is not released yet. I will change the script to use 9k topics and retry the test

Sorry, yes the 2.12 hasn’t be released yet. If your goal is to measure the performance of the cluster handling load across 10K+ topics, then one option is to run multiple versions of the same test that use a different topic and namespace, e.g. 2 simultaneous benchmark tests of 5K topics configured to use different topics.

Thanks <@U9NHC1ETG>. is there any limitation on number of tenants? we are planning to have 30k tenants. is that okay?

There aren’t any limitations that I am aware of.