Problem: impossible to know the order of key insertion

[yrashk]

There are some cases when knowing the order of key insertion would be beneficial.

For example, replication that can be started and stopped at any time. The replica would only have to provide the last seen key and the master can provide the immediately following keys.

Proposed solution 1: unconditionally journal keys (HLC key) in a separate database in the same database environment and provide a set of instructions to navigate these. Downside: complexity and unconditionality (can't opt out).

Proposed solution 2: add ASSOC/JOURNAL instruction that will journal the associate the key value pair and journal the key. Downsides: leaks the journal to the namespace, makes it impossible to enforce this across the entire database, meaning at least the replication case won't benefit from it.