Support dual-labeled metrics in structured_missing_columns

[BenWu]

Description

The join in the structured_missing_columns view doesn't account for dual-labeled so the columns are always marked as not existing in the schema.

Reviewer, please follow this checklist

[github-actions]

This PR extends the udf.remove_label_from_metric_path UDF to also match metrics.dual_labeled_ paths (in addition to metrics.labeled_), so dual-labeled metric columns are truncated to their first three path segments and can join against INFORMATION_SCHEMA.COLUMN_FIELD_PATHS in the structured_missing_columns view. A test case for a dual_labeled_counter path is included.

The logic is sound: both labeled and dual-labeled paths reduce to the first three segments regardless of label count, which matches the schema column name and the view's join. The prefix check generically covers all dual-labeled metric types. I have no findings to raise.

[scholtzan]

Integration report

• View full diff, if available (SQL + DAGs) (requires private-bigquery-etl access)
• Private CI run
• Bigeye monitoring dry-run: ✅ plan clean — view plan (requires private-bigquery-etl access)