Creates ADR with conventions for zod schemas

[Rotorsoft]

Link to Issue

Closes: #8823

Description of Changes

• Adds new ADR

Next Steps

• Once this is approved, we'll create a new cleanup bucket to enforce these conventions

[rbennettcw]

I think it's not a bad idea to have an alias type for the timestamps, e.g. TIMESTAMP_DATE that reduces the cognitive load of needing to remember the pattern + provides flexibility to change/refine the type in the future for whatever quirks may arise

[Rotorsoft]

I think it's not a bad idea to have an alias type for the timestamps, e.g. TIMESTAMP_DATE that reduces the cognitive load of needing to remember the pattern + provides flexibility to change/refine the type in the future for whatever quirks may arise

We had zDate, but I don't think it adds a lot of value by just coercing to date... probably better to remember to coerce all dates explicitly in the schemas.

PG_INT is different because it enforces PG integer limits