How do configure scaling based on latency in AWS E

Question

Latency of my application deployed to AWS Elastic Beanstalk is generally about 50 ms, I want another instance to start if average latency goes to 500ms (1/2 second). You can change the scaling Trigger in the configuration and select Latency, but the smallest Unit of Measurement is Second and Upper and Lower Scale Increments values must be an integer or percent.

So it doesn't seem possible to set a trigger as required.

Answer 1

The properties "Lower breach scale increment" and "Upper breach scale increment" is the number of EC2 instances to add or remove. This has to be an integer or a percentage.

http://docs.aws.amazon.com/elasticbeanstalk/latest/dg/command-options.html

For using subsecond latencies as triggers for scaling, set the values of UpperThreshold and LowerThreshold to "0.5" for 500 ms. The Unit of Measurement will be Second.