feat(coordinator): use monotonically increasing boot IDs

[YuMo12268]

No description provided.

[YuMo12268]

@un-lock-able Could you please review this PR? Thanks!

[un-lock-able]

Do we really need to store boot_id inside database or we can just store it in a local file? May need to discuss with @BobAnkh .

[BobAnkh]

Do we really need to store boot_id inside database or we can just store it in a local file? May need to discuss with @BobAnkh .

We will discuss this next week.

[un-lock-able]

We can switch to using UUIDv7 that includes a timestamp in itself, so that the coordinator doesn't have to store the last boot_id in a DB or in a file. Upon seeing a UUID that is larger than the last seen UUID, the agent can assume that the coordinator has restarted, reconnect to it, and perform a sync. If the UUID is smaller than the last seen UUID, however, the agent can assume that some unexpected state has occurred and exit directly rather than attempting to recover from an ambiguous state.