Inconsistent error messages in case if recommendation is not possible

[sancyx]

If you invoke scaleOut with the below payload, which is the exact cluster layout returned by recommendation call for same resource number, you get an error 400 (could not recommend cluster with the requested resources). However if you set "sumNodes": 1 for t3.medium spot pool you will get error 500 (error: &errors.errorString{s:\"there's already enough resources in the cluster. Total resources available: CPU: 4, Mem: 8\"}). In case you decrease desiredMem = 3 you'll get the same layout recommendation. It would desirable to have same behaviour in each case, either return the same clusterLayout or return the same error code and message, since in each case you have enough resources in your cluster.

{
    "desiredCpu":2,
    "desiredMem":4,
    "desiredGpu":0,
    "onDemandPct":80,
    "actualLayout": [
        {
          "instanceType": "t3.medium",
          "vmClass": "regular",
          "sumNodes": 1
        },
        {
          "instanceType": "t3.medium",
          "vmClass": "spot",
          "sumNodes": 0
        },
        {
          "instanceType": "t2.small",
          "vmClass": "spot",
          "sumNodes": 0
        },
        {
          "instanceType": "t2.medium",
          "vmClass": "spot",
          "sumNodes": 0
        },
        {
          "instanceType": "t2.large",
          "vmClass": "spot",
          "sumNodes": 0
        },
        {
          "instanceType": "t3.large",
          "vmClass": "spot",
          "sumNodes": 0
        },
        {
          "instanceType": "m4.large",
          "vmClass": "spot",
          "sumNodes": 0
        }
    ]
}```

[waynz0r]

I would not treat there's already enough resources in the cluster as an error.