Parameterize the target response rate alarm threshold

[aaronxsu]

We recently added an alert on the ALB target response rate to monitor the app server latency since the running instance on production has scaled down to 1, and it is beneficial for us to have visibility of latency to take actions just in case.

The threshold is currently hard-coded in the terraform code. We should parameterize it and make it configurable through tfvars.

This issue is to:

• update the current threshold to a reasonable level for the prod alarm in the console. this should be done first since the following two won't take effect until the next release.
• parameterize the target response rate alarm threshold
• update the tfvars