Scaling experiments memory

Hi, I am trying to run a few experiments to understand scale better for my workload. One of the questions I had was how much memory is too small to understand scaling.

Let's say I have 3 nodes (running ES 2.3) each with 1GB RAM on 10GB SSD's and with about 2.5 million docs and 1 replica, I've noticed that the JVMMemoryPressure stays at around 70% which seems rather high. The problem is with this load the queries are taking much longer to complete i.e. much more latency.

Question: Does anyone have any pointers on how I should approach determining what is causing this latency?

I am guessing it's the memory pressure, but I am not sure how to understand the node stats metrics to see why there is so much JVMMemoryPressure.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.