We have a web server on which we are going to run several applications. All of them will share the database servers and memcached, but each application has its own mySQL database and all memcached keys for each application has a prefix.
Possible scenario:
If the memcached server in our cluster begins to develop, we want to automatically notify someone (operational system administrator) via email / iphone push or any other appropriate way.
If we are going to install 150 identical applications for our clients on our servers, and the memcached server dies, all 150 applications will individually figure this out and contact our system administrator, which will certainly think about getting a new job in which he or she does not going to wake up, having received 150 messages sent 4:15 in the morning.
Possible Solution:
One idea is to configure an external server to handle errors, which receives the sent $ _POST or cURL request, and processes the storage of the error message depending on the severity of the actual error message. Of course, when receiving a call about an error, it will be checked that if the same memcached server was already sent offline, there would be no need to spam the system administrator with additional reminders ...
:
!