Topic
No replies
ShwetaSSharma
ShwetaSSharma
2 Posts
ACCEPTED ANSWER

Pinned topic Removel of instance is taking too long time for response based scaling policy

‏2013-04-19T05:42:04Z |

Hi,

I am working on "Response based Scaling policy" in which I am able to do scale in and out when the threshold condition is satisfied. But when the threshold condition is satisfied for scale out, time taken to remove an instance is too long almost 3hours. What do we need to change to remove the instance in short span of time? Below is the template for my response based policy :

{
    "scaling": {
        "triggerEvents": [
            {
                "metric": "Tomcat_ResponseTime.server_response_time",
                "scaleOutThreshold": {
                    "value": 500,
                    "type": "CONSTANT",
                    "relation": ">="
                },
                "conjection": "OR",
                "scaleInThreshold": {
                    "value": 111,
                    "type": "CONSTANT",
                    "relation": "<="
                }
            }
        ],
        "min": 1,
        "max": 2,
        "triggerTime": 60,
        "role": "Tomcat"
    },
    "name": "Tomcat-tomcat",
    "packages": [
        "TOMCAT",
        "TOMCAT_SCRIPTS"
    ],
    "roles": [
        {
            "parms": {
                "ARCHIVE": "https://127.0.0.1:9444/storehouse/user/deployments/d-810d72ee-b3b9-46ae-8e3b-f72d306149b1/artifacts/ZipTest.war",
                "triggerTime.ResponseTimeBased": 100
            },
            "type": "Tomcat",
            "name": "tomcat"
        }
    ]
}

 

However, In case of CPU based Scaling policy, It does not take much time to add/remove machine.

Kindly Revert.