Hi William, I have performed a small test: created 10000 queues using bash script which calls Joloka API using curl in a loop. It works a bit slow because each API call needs to perform an authentication and to create a session. Another method is to auto-create queues when attempting to consume or browse each queue. The memory is allocated for each object (even if it is not used). In this case, Old Gen heap increased to 900 Mb, and after restart it has been decreased to 750 Mb. The restart took about 30 seconds. There is an address setting which may help to reduce memory footprint for each queue - initial-queue-buffer-size (added in 2.37.0).
The most important problem is that a large number of objects slows down each start, restart or failover of the broker, affecting its availability. ср, 5 мар. 2025 г. в 21:35, William Crowell <wcrow...@perforce.com.invalid>: > Good afternoon, > > Are there any limitations or concerns with creating millions of topics and > queues within Apache Artemis that have low volumes of messages in each > topic and queue? I do not think there are as I believe Artemis should be > able to handle this use case. Important question is: Is it scalable? > > Regards and have a great day, > > William Crowell > > > This e-mail may contain information that is privileged or confidential. If > you are not the intended recipient, please delete the e-mail and any > attachments and notify us immediately. > >
