Some of the sample values for auto-scaling thresholds doesn't make sense.

[Nirmal Fernando]

https://cwiki.apache.org/confluence/display/STRATOS/4.0.0+Sample+Auto-scaling+Policy+Definition#id-4.0.0SampleAuto-scalingPolicyDefinition-loadThresholds

* memoryConsumption - is a percentage and cannot go beyond 100 since a
process cannot take more than total physical memory.

formula: (usedMemory / totalMemory) * 100

* loadAverage - is a percentage and it can go beyond 100% but 6000 is
not a practical value.

formula: (loadAvg/cores) * 100

* Further you do not want to auto-scale when the gradient is 0 (which
means 'y' doesn't change with time (x)).

Please see following sample:

{
  "id": "autoscale-policy-1",
  "loadThresholds": {
    "requestsInFlight": {
      "average": 50,
      "secondDerivative": 0.2,
      "gradient": 10,
      "scaleDownMarginOfGradient": 20,
      "scaleDownMarginOfSecondDerivative": 0.5
    },
    "memoryConsumption": {
      "average": 70,
      "secondDerivative": 0.2,
      "gradient": 10,
      "scaleDownMarginOfGradient": 20,
      "scaleDownMarginOfSecondDerivative": 0.5
    },
    "loadAverage": {
      "average": 100,
      "secondDerivative": 0.2,
      "gradient": 20,
      "scaleDownMarginOfGradient": 20,
      "scaleDownMarginOfSecondDerivative": 0.5
    }
  }
}

@Mari, would you be able to do these amendments?

Thanks.

-- 
Best Regards,
Nirmal

Nirmal Fernando.
PPMC Member & Committer of Apache Stratos,
Senior Software Engineer, WSO2 Inc.

Blog: http://nirmalfdo.blogspot.com/