Answer may be:

If one example crashes, all of them go offline, and when the missing information is replayed as a customer makes a query this will place a greater burden on the computer system. To prevent this, if the software has been designed to mitigate database pileups, it will have a marginal effect.

Another approach is to pull up an example of Memcached using the Internet addresses of the missing computers on a new computer.

The software is another way to mitigate server outages as it allows you the freedom with limited effort to modify the Memcached server list.

Another alternative that certain Memcached customers introduce for Memcached server failure is establishing the load balancer. The application will continue to attempt to submit a re-queue while the Memcached database goes offline.