Environment health has been unstable, fluctuating between “warning”,“degraded” and “OK” (with constituting instances [autoscaling with a minimum of 3 instances ] periodically changing between these states) , the cpu utilisation is low and i have looked in the logs but haven’t found any probable cause.
x % (on the average less than 9%) of the requests are failing with HTTP 5xx.
Any ideas why this happens or any troubleshooting tips?
It may not solve the problem immediately but it’s probably worth upgrading the instance type to m4.large (similar specs, cheaper than m1 equivalent) or similar. The newer generations will give you more compute (ECU), newer CPUs and m1 => m4 will also give you SSD storage as well.