I am wondering if I am missing something here but there seems to be a fundamental flaw with the front end/back end web services topology...
basically if there is an issue with e.g. the EWS website on a MBX server hosting a copy of my mailbox I would be unable to set my out of office or have other EWS related issues...However the HLB monitoring and directing traffic to my CAS boxes would continue to do so as there is no issue with EWS on the front end, the CAS would proxy the traffic to the back end EWS website and have a problem and the very nature of high availability and automatic failover here has failed. The EWS issue on the mailbox server IIS site is outside of the scope of the DAG to notice so my mailbox would be staying on the server with the EWS back end site issue until I noticed and or moved it...

Are my statements and understanding co

In theory, Managed Availability would recognize the problem and ultimately reboot the server in an attempt to fix it.

Otherwise, it would be on the admin to take action such as activating the database to another server or rebooting the server in response to alerts or end-user complaints.

So in this case is this not making the high availability less available...was moving to the front end/back end web services from CAS side rendering worth it?

So in this case is this not making the high availability less available...was moving to the front end/back end web services from CAS side rendering wor

yes but what issues could there have been on the mailbox server? there was no back end IIS sites like there is now...anything that failed on mbx 2010 would cause a DAG failover...I am talking about the high availability of web services and they are surely less resilient with the front end back end model as the load balancers can only health check traffic direction to CAS servers for web services not to the back end.

thoughts?

yes but what issues could there have been on the mailbox server? there was no back end IIS sites like there is now...anything that failed on mbx 2010 would cause a DAG failover...I am talking about the high availability of web services and they are surely less resilient with the front end back end model as the load balancers can only health check traffic direction to CAS servers for web services not to the back end.

tho

If you are doing a healthcheck probes with your 2013 load balancer per the link below, then the load balancer sholdnt be sending to any server that cant handle the request correctly

http://blogs.technet.com/b/exchange/archive/2014/03/05/load-balancing-in-exchange-2013.aspx

Managed Availability will attempt to fix the issue or reboot the server so the database is moved to another server. There are always compromises, but I dont see how that makes 2013 less available.