Cache Engine Versions

Correct Answer: New version has a different performance profile

A new version might have a different performance profile, leading to increased latency. The other responses are possible but less likely explanations. Incompatible configurations would likely result in an error, rather than latency, and the upgrade process itself is separate from the observed latency.

Correct Answer: Upgrade to a more recent version

Because the performance issue is traced to a known regression in the current version and the vendor explicitly states that the next stable patch fixes it without breaking changes, upgrading to that newer stable release is the correct and durable remediation. Downgrading reintroduces older bugs and is unnecessary when a fix exists. Restarting may provide temporary relief but does not resolve the underlying defect. Upgrading to an experimental version introduces risk and is inappropriate for production.

Correct Answer: Create a new ElastiCache for Redis cluster/replication group and explicitly select the required older engine version at creation (or restore from a snapshot taken on that version).

In ElastiCache for Redis, the engine version must be chosen when you create the cluster/replication group, or you can restore from a snapshot that was taken on the required version. ElastiCache does not support engine downgrades, so launching on a newer version and downgrading later is not possible. Parameter groups control runtime configuration and do not determine the engine version. Cloning or modifying an existing cluster does not change the engine version unless you are restoring from a snapshot that already has the older version. AWS Support does not enable deprecated engine versions on demand.