This page describes how your Memorystore for Memcached instance behaves during a version upgrade operation, how an upgrade operation can affect your application, and when you should run an upgrade operation. To learn how to upgrade an instance's Memcached version, see Upgrade the Memcached version of an instance.
Upgrading an instance has performance and storage implications for your application.
Best practices for upgrading an instance's Memcached version
Upgrade your instance during periods of low instance traffic to minimize the impact of the full cache flush caused by the upgrade operation. To learn how to monitor instance traffic, see Monitor Memcached instances.
Version upgrade behavior
The Memcached instance undergoes a complete cache flush during a version upgrade operation. Multi-node instances are partially available during the upgrade operation. You cannot downgrade the instance to an earlier version once it has been upgraded.
An upgrade operation also causes the nodes to undergo a short connection reset of a couple minutes, or less. Applications should incorporate retry logic in the code to be able to reconnect to the nodes. The discovery endpoint and memcached node endpoints remain the same.
Because of the short term connection break, there may be a small amount of stale or inconsistent data that didn't write or update to the cache for the short amount of time the Memcached node was unavailable.