Might not be dash but it's definitely something on frdev1001: https://grafana.wikimedia.org/dashboard/db/fundraising-host-overview?refresh=5m&panelId=13&fullscreen&orgId=1&var-server=frdev1001.frack.eqiad.wmnet&var-datasource=frack.codfw%20prometheus&from=now%2Fw&to=now
Switch the dropdown to frdb100[12] (write and read servers) to see the correlation
It seems likely that this is the cause of intermittent "too many processes" alerts on frdb1001