With AWS Elastic Beanstalk, I have an app setup with enhanced health monitoring. However, the app uses long polling so that it can sync between a mobile app and the web app. The long poll is a 60 second timeout, and this causes enhanced health to detect the 60 second latency and determine that the health of the instance is degraded. It will then replace the instance with a new one.
If I use Basic health monitoring, AWS apparently doesn't use the latency as a determining factor for health.
Is there a way to use enhanced health with a long polling app that will avoid the false positives of long latency?