Hi Team, We have the max-size-bytes configured as -1 in the address-setting and global-max-size as 1/5th of the broker xmx. Recently we had a situation of DLQ with more messages, which caused paging of all addresses affecting broker performance. To avoid such situation, we want to configure max-size-bytes for DLQ with a lower value so that DLQ messages do not cause paging of other addresses. Can you clarify below
* How is address size calculated? * When messages are sent to broker, are the messages present only in memory or they are also written to disk (before paging begins)? * Are only paged messages written to disk? * How is address size different from Queue persistent size? * Why is address size very less compared to Queue persistent size? How to calculate the max-size-bytes for a queue based on expected persistent size for the queue? * How to ensure memory is not unnecessary held by messages in DLQ/any custom error queue? Thanks, Mohanavalli A
