Resource group is not working as expect

[njalan]

Below is my resource group file.
Sub group opn is for report which is running slow and it takes 30 seconds by 5 users.
Sub group dss is for report which is running fase and it takes 2 seconds by 100 users. Concurrency is very high.
Report running by subgroup opn is more import then report running by group dss.
My point is that I want to make sure Sub group *opn at any time has 20% of the resource but actually when there are high workload for Sub group dss then *opn is also affected.
How can I make sure Sub group opn has the high priority then dss and make sure it always has 20% of the resource. I just want to make sure the reports running by sub group opn always get enough resource to finish.

{
"rootGroups": [
{
"name": "global",
"softMemoryLimit": "90%",
"hardConcurrencyLimit": 80,
"maxQueued": 300,
"jmxExport": true,
"subGroups": [
{
"name": "zlb",
"softMemoryLimit": "60%",
"hardConcurrencyLimit": 20,
"maxQueued":100
},
{
"name": "opn",
"softMemoryLimit": "20%",
"hardConcurrencyLimit":20,
"maxQueued": 40
},
{
"name": "dss",
"softMemoryLimit": "20%",
"hardConcurrencyLimit":30,
"maxQueued": 160
}
]
}
],
"selectors": [

{
  "user": "xxxx",
  "group": "global.opn"
},
{
  "user": "yyy",
  "group": "global.dss"
},

{
  "user": ".*",
  "group": "global.zlb"
}

],
"cpuQuotaPeriod": "1h"
}

[hashhar]

Resource groups as they exist today only provide admission control. Once a query is selected for execution by resource groups they no longer have any control i.e. they cannot limit amount of resources a group can use at runtime, they can only prevent future queries to run in a group if a group has exceeded it's limits.