fix(gold): resolve task_status_intervals name collision

src/ingestion/gold/schema.yml

@@ -359,7 +359,7 @@ models:
         tests:
           - not_null
 
-  - name: task_status_intervals
+  - name: task_status_spans
     description: >
       Per-issue status spans for the task-delivery observation pipeline: each
       status event paired with the next; the last span ends at the close for

src/ingestion/gold/task_metric_observations.sql

@@ -17,7 +17,7 @@
 -- measure is emitted through macros/metric_observation_measures.sql.
 --
 -- The per-issue reconstruction is materialized upstream (task_issue_state,
--- task_status_intervals): ClickHouse re-inlines every WITH reference, so the
+-- task_status_spans): ClickHouse re-inlines every WITH reference, so the
 -- measure branches would otherwise re-run it once per branch.
 --
 -- Grain: event rows per closed issue for the median metrics (dev_time_hours,
@@ -43,7 +43,7 @@ issue_state AS (
 ),
 status_intervals AS (
     SELECT *
-    FROM {{ ref('task_status_intervals') }}
+    FROM {{ ref('task_status_spans') }}
 ),
 -- Per closed issue (metric_date = close date): dev seconds, lead, pickup.
 -- Only spans started before the close count — live rework on a reopened

src/ingestion/gold/task_status_intervals.sql → src/ingestion/gold/task_status_spans.sql

@@ -3,7 +3,7 @@
     engine='MergeTree',
     order_by=['insight_source_id', 'issue_id', 'interval_start'],
     schema='insight',
-    alias='task_status_intervals',
+    alias='task_status_spans',
     tags=['gold'],
     query_settings={
         'max_memory_usage': 1610612736,