fix(workflow_engine): Reduce DataCondition cache TTL

[kcons]

7d is much longer than we need, and means bad cache issues need manual intervention.
4h should be basically just as good for performance purposes, and seems more reasonable.
While here, make the TTL required in the API we're using so it's clear where things are cached for long durations.

[sentry]

Bug: The caching logic in CachedFunction.__call__ incorrectly treats a cached None value as a cache miss, causing the decorated function to be re-executed unnecessarily.
_{Severity: MEDIUM}

Suggested Fix

Use a sentinel object to distinguish a cache miss from a stored None value. Initialize a unique object, e.g., _sentinel = object(), and use it as the default for cache.get. The subsequent check should be against this sentinel object, not None.

Prompt for AI Agent

Review the code at the location below. A potential bug has been identified by an AI
agent. Verify if this is a real issue. If it is, propose a fix; if not, explain why it's
not valid.

Location: src/sentry/utils/function_cache.py#L89-L90

Potential issue: The caching mechanism in `CachedFunction.__call__` uses `cache.get(key,
None)` and then checks `if result is None` to determine if there was a cache miss. This
logic is flawed because it cannot distinguish between a cache miss and a case where the
value `None` was legitimately cached. As a result, any cached function that returns
`None` (such as `_load_service_hook`) will have its result recomputed on every call
instead of being retrieved from the cache, defeating the purpose of caching for these
specific return values.

_{Did we get this right? 👍 / 👎 to inform future reviews.}