On the standby, we can skip certain sanity checks, such as verifying whether a key already exists. These checks are unnecessary on the standby because they are performed on the primary before the data is replicated. Skipping these checks helps optimize the performance, as verifying the existence of a key involves reading from the stable storage table and performing a lookup, which can be expensive.

To enhance efficiency, the standby can operate under the assumption that the data it receives is already valid and consistent as handled by the primary.