Collect exportable engine metric representing worker capacity usage for auto scaling #1465
Description
Describe the desired outcome from the user's perspective
As devops, I want to define an auto-scaling policy based on engine capacity usage.
Acceptance criteria
- Describe usage as a percentage of available capacity
- Auto-close new TCP connections at full capacity instead of unbinding completely
Additional context
Intention of auto-close at capacity is to prevent undesirable side effects when instance appears unhealthy.
Engine worker capacity metric units should be suitable for AWS AutoScaling via CloudWatch custom metric.