feat: Add force_destroy options to prevent accidental data loss

[Adam-D-Lewis]

Summary

This PR implements the solution for #8 by adding configuration options to prevent accidental deletion of storage buckets and database credentials when disabling the MLflow plugin.

Changes

• Added force_destroy_storage and force_destroy_db_creds configuration options to the MLflow plugin schema
• Both options default to false to preserve resources by default
• Updated AWS, GCP, and Azure Terraform configurations to respect these options
• Added lifecycle prevent_destroy rules to storage resources
• Configured database credentials to be preserved unless explicitly destroyed

Configuration

Users can now control resource deletion with explicit flags:

mlflow:
  enabled: false
  force_destroy_storage: true  # defaults to False
  force_destroy_db_creds: true # defaults to False

Test plan

• Python configuration validates correctly
• Test on AWS deployment with MLflow enabled/disabled
• Test on GCP deployment with MLflow enabled/disabled
• Test on Azure deployment with MLflow enabled/disabled
• Verify storage buckets are preserved when force_destroy_storage: false
• Verify database credentials are preserved when force_destroy_db_creds: false
• Verify resources are destroyed when flags are set to true

Fixes #8

🤖 Generated with Claude Code