Exercise 1: Choosing Redundancy for Both Zone and Regional Protection — Possible Solution
====================================================================

The requirement described has two real, distinct parts: protection
against a single zone outage, AND protection against a full regional
disaster. Since the target region does support Availability Zones,
every redundancy option is genuinely available to choose from.

LRS only protects against drive/server/rack failure within a single
datacenter -- it would not survive an entire zone becoming
unavailable, let alone a regional disaster. ZRS solves the zone-outage
problem (replicating across 3+ real Availability Zones), but stays
entirely within the primary region -- a full regional disaster would
still take out every zone-redundant copy at once, since none of them
exist outside that region. GRS/RA-GRS solves the regional-disaster
problem (replicating to a real, geographically distant secondary
region), but replicates using LRS within the primary region itself --
meaning a single zone outage in the primary region isn't specifically
protected against the way ZRS protects against it.

ANSWER: GZRS (or RA-GZRS, if read access to the secondary region
during an outage is also needed) is the right recommendation. It
combines ZRS's own real zone-level protection in the primary region
with GRS's own real geographic replication to a secondary region --
the only option among the four that genuinely protects against both
described failure scenarios at once, rather than only one of them.

WHY THIS WORKS AS AN ANSWER
------------------------------
This correctly evaluates each of the four real options against BOTH
stated requirements individually, showing exactly which specific
failure scenario each option fails to cover, rather than simply
picking the "highest number of nines" option without explaining why
the other high-durability option (GRS) still falls short of the
zone-protection requirement specifically.