Documentation: Clarify in-memory state behavior and persistence recommendations

## Summary
When developing Azure Functions with in-memory storage (e.g., using global variables or Maps for caching/state), the data is lost when:
- The function app restarts
- The function scales to new instances
- During local development when `func start` is restarted

This can lead to confusing behavior during development and testing, where data created in one session is not available after restart.

## Reproduction Steps
1. Create a function that stores data in a global variable/Map
2. Call the function to store some data
3. Restart `func start` (locally) or trigger a function app restart (in Azure)
4. Attempt to retrieve the previously stored data
5. Data is not found (404/null)

## Suggestion
Consider adding documentation or warnings that:
1. Clarify that in-memory state is ephemeral and instance-specific
2. Recommend persistent storage options (Azure Table Storage, Cosmos DB, Redis Cache) for any data that needs to survive restarts
3. Perhaps emit a warning in the Functions runtime when global state modifications are detected, suggesting persistent alternatives

## Environment
- Azure Functions Core Tools: 4.6.0
- Runtime: Node.js v4 programming model
- OS: macOS

## Related
This is a documentation/developer experience issue, not a bug. The behavior is expected, but could be better communicated to developers new to serverless.