This one is becoming more and more urgent for us, as we see a lot of
Beanstalk implementations lately.
At start it should not be very fancy. For example, just a check on
Environment Health of each Beanstalk environment would be a very big help.
Basically it's just a Cloudwatch Metric in
the AWS/ElasticBeanstalk namespace,
called EnvironmentHealth. If that value is above 15, something is
wrong and an alert needs to be sent.
Or even a way to create custom Cloudwatch metrics based monitors
would be a start.