We are using Enterprise 6.0.
We have a index node + a replica with a large set of indexes.
We are watching the total mutations remaining stat climb and climb (with any query requiring consistency timing out) and looking in dismay at the drain rate. The mutations in the tens of thousands and growing and the drain rate is maybe in the teens when it isn’t zero.
I can find nothing at all on line instructing about the impact of this phenomenon nor how to go about increasing the rate.
What is the performance significance of this and how can we mitigate. Is this a function of how many indexes we have per node? Can we get some guidance.
I’m also having difficult reading the stats.
I navigate into a index node and inspect the stats provided per index. This is where I see the ever growing total remaining mutations figure and anemic drain rate. The number are the same for all the indexes so I assume it’s an index node figure. Is this correct?
The ram profile for the server looks great: 30 % of RAM used and north of 100G of Index RAM remaining.