Permanently and irreversibly delete a holdout by its unique ID. This is a destructive operation that cannot be undone.

⚠️ CAUTION - Permanent Deletion:

• All holdout configuration data will be permanently lost
• Associated metric data will be deleted from MetricsHub
• Experiment rules and historical data will be removed
• There is no way to recover this data after deletion

What gets deleted:

• Holdout configuration (traffic allocation, audience conditions, etc.)
• All associated experiment rules and their settings
• Metric tracking data and historical performance
• System metadata and lifecycle timestamps

Safer Alternative - Archiving: If you want to stop a holdout without permanent data loss, consider archiving instead:

1. Update the holdout status to concluded (if currently running)
2. Use the PATCH endpoint to set archived: true

Archived holdouts preserve all historical data while removing them from active lists and preventing further modifications.

When to use DELETE vs ARCHIVE:

• DELETE: Use only when you're certain the holdout and all its data should be permanently removed
• ARCHIVE: Recommended for most cases - preserves data for analysis while cleaning up active experiments

Important Notes:

• Can delete holdouts in any status (draft, running, concluded)
• Operation affects both database and MetricsHub systems
• This corresponds to the delete_url link in holdout responses