Bug: ansible_name column has no uniqueness constraint or input validation

[mrveiss]

Problem

The ansible_name column added in PR #1932 (#1814) has no uniqueness constraint in the database model and no min_length validation in the Pydantic schema.

Uniqueness

Two nodes could have the same ansible_name, causing Ansible --limit to target the wrong node (or both). Since ansible_name is used for --limit targeting and Ansible inventory host keys, duplicates would cause silent misrouting of playbook executions.

Validation

NodeUpdate.ansible_name accepts any string including empty string "". While "" is falsy in Python's or operator (so the ansible_target property falls back to IP), an explicit Field(min_length=1) would prevent confusion.

Discovered During

Implementation of #1986 (ansible_name auto-populate from heartbeat)

Fix

1. Add UniqueConstraint or unique index on ansible_name (nullable unique — NULL is allowed, duplicates are not)
2. Add Field(min_length=1) to NodeUpdate.ansible_name and NodeCreate.ansible_name
3. Add migration for the unique constraint

Impact

Severity: medium — Silent misrouting of Ansible commands if duplicates exist.

[mrveiss]

Fixed in PR #2352 (merged). Added unique constraint + partial index migration on ansible_name, min_length=1 validation, IntegrityError handling in create/update/heartbeat endpoints, extracted _auto_populate_ansible_name helper.