The controller kept one report per machine, replaced, so a resource nothing can ever apply looked like a failure that had just happened, every few minutes, for ever. It now counts identical reports and status says stuck after three.
jschoubben