I just wanted to ask for advice and to understand if anyone has experienced anything like this.
We have received three heap dumps from healthy nodes hosted on AWS. The one which had the largest heap usage (1.7G) showed that 1G of the heap was consumed by declared pages. There were a total of 1,992 declared pages, out of which 1,697 shared the same name pattern.
We don't use any complicated reference data design patterns, so we wanted to understand if this was a product issue. This doesn't sound like a lot off declared pages, this seems a typical amount, so we're unsure as to why the size in memory is so large.
These appear to be node level parameterized data pages. Is this correct? Are they list or single page?
If they are lists, the size issue would not come from the number of data pages but, rather, the length and width of each data page. In addition, persistence will be handled based on the refresh strategy associated with each data page.