These servers are due for a refresh this FY. They have fallen far enough behind we have had to switch morelike traffic from the eqiad cluster to codfw to keep these older servers from maxing out CPU during the busiest parts of the week and effecting user latency.
The server spec we are using these days for elasticsearch nodes is higher than these old nodes. Ideally we should match specs with elastic10{17..31}.eqiad.wmnet. The refresh budget on these servers is not enough for all of them, due to the higher spec, as such the discovery capex budget will be used to cover any additional funds required.