Looking at this a little deeper, it looks like the pod manager itself should be returning some better error messages. If the system with the most available memory has 8GB and I request 9, the error message returned by the podm API just says "There are no computer systems available for this allocation request", with no mention of the reason being a shortage of memory. I don't think we'll be able to fix this in Valence until the pod manager is improved here.
Looking at this a little deeper, it looks like the pod manager itself should be returning some better error messages. If the system with the most available memory has 8GB and I request 9, the error message returned by the podm API just says "There are no computer systems available for this allocation request", with no mention of the reason being a shortage of memory. I don't think we'll be able to fix this in Valence until the pod manager is improved here.