Use MIN/MAX/AVG for sp_QuickieStore wait-time column aggregations

The plan-level wait-stats aggregation in the #query_store_wait_stats
insert used SUM() for every value including the ones typed and named
as min_query_wait_time_ms and max_query_wait_time_ms. Summing per-interval
minima and maxima inflates the output by the number of captured intervals
and produces numbers the column names promise they won't.

Changed the min_/max_ columns to MIN()/MAX(). Also changed
avg_query_wait_time_ms from SUM() to AVG(); summed averages are the
same kind of unit error. The outer GROUP BY is per (plan_id,
wait_category_desc), so MIN/MAX/AVG land on the right scope.

Left the HAVING SUM(min_query_wait_time_ms) > 0 as-is — as a gate for
"any wait time recorded in any interval" it's correct and changing it
would alter the row-filter semantics.

total_query_wait_time_ms still uses SUM — that one matches its name.

Verified the sproc installs and returns correctly-shaped wait output
against PerformanceMonitor on SQL Server 2022.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: erikdarlingdata

sp_QuickieStore/sp_QuickieStore.sql

@@ -10832,13 +10832,13 @@ SELECT
     total_query_wait_time_ms =
         SUM(qsws_with_lasts.total_query_wait_time_ms),
     avg_query_wait_time_ms =
-        SUM(qsws_with_lasts.avg_query_wait_time_ms),
+        AVG(qsws_with_lasts.avg_query_wait_time_ms),
     last_query_wait_time_ms =
         MAX(qsws_with_lasts.partitioned_last_query_wait_time_ms),
     min_query_wait_time_ms =
-        SUM(qsws_with_lasts.min_query_wait_time_ms),
+        MIN(qsws_with_lasts.min_query_wait_time_ms),
     max_query_wait_time_ms =
-        SUM(qsws_with_lasts.max_query_wait_time_ms)
+        MAX(qsws_with_lasts.max_query_wait_time_ms)
 FROM
 (
     SELECT