|
Our filtering implementation is very CM-specific and will change every release. I think we need to solve this in a more general fashion. I'm not sure exactly how but if /v1/state-overview could take a additional constraints in the form of a subquery, could that work?
In an ideal world this endpoint could apply the filter criteria we specify directly in the DB, but at the scale we're talking about it isn't unfeasible to ask the /v4/nodes endpoint for the statuses and calculate the aggregates ourselves. That would depend on PDB-1820. I'd put that ticket about this one, and put this one on the long finger for a time when we're sure we have a performance problem.
|