Plan değişimi P443

(@condition nvarchar(4000),@incTest bit,@types1 nvarchar(4000),@sid int,@dbName nvarchar(4000),@tag nvarchar(4000),@top int)SELECT TOP (@top) c.ChunkId, c.DocId, d.SourceType, d.SourceRef, d.Title, c.Heading, c.Text, CAST(ft.[RANK] AS float) AS Score
FROM CONTAINSTABLE(rag.KbChunkText, Text, @condition) AS ft
JOIN rag.Chunk AS c ON c.ChunkId = ft.[KEY]
JOIN rag.Document AS d ON d.DocId = c.DocId
WHERE (d.IsActive = 1 OR (@incTest = 1 AND d.ExclusionReason = 'TestDb')) AND d.SourceType IN (@types1)
AND (@sid IS NULL OR d.ServerID = @sid) AND (@dbName IS NULL OR d.DatabaseName = @dbName)
AND (@tag IS NULL OR EXISTS (SELECT 1 FROM STRING_SPLIT(d.Tags, ',') AS t WHERE LTRIM(RTRIM(t.value)) = @tag))
ORDER BY ft.[RANK] DESC, c.ChunkId

Plan change case #P443 | DB: DB-11 | Regression | Confidence: Low (review pending)
Query: (@condition nvarchar(?),@incTest bit,@types1 nvarchar(?),@sid int,@dbName nvarchar(?),@tag nvarchar(?),@top int)SELECT TOP (@top) c.ChunkId, c.DocId, d.SourceType, d.SourceRef, d.Title, c.Heading, c.Text, CAST(ft.[RANK] AS float) AS Score FROM CONTAINSTABLE(rag.KbChunkText, Text, @condition) AS ft JOIN rag.Chunk AS c ON c.ChunkId = ft.[KEY] JOIN rag.Document AS d ON d.DocId = c.DocId WHERE (d.IsActive = ? OR (@incTest = ? AND d.ExclusionReason = ?)) AND d.SourceType IN (@types1) AND (@sid IS NULL OR d.ServerID = @sid) AND (@dbName IS NULL OR d.DatabaseName = @dbName) AND (@tag IS NULL OR EXISTS (SELECT ? FROM STRING_SPLIT(d.Tags, ?) AS t WHERE LTRIM(RTRIM(t.value)) = @tag)) ORDER BY ft.[RANK] DESC, c.ChunkId
Change: 2026-10-08 09:00 UTC, plan 66753 → 70259, pattern: Transition
Impact (Query Store, per execution): duration 17.8 → 26.1 ms (+46%); CPU 16.8 → 25.9 ms (+54%); reads 2,401 → 3,902 (+63%); rows 8.76 → 15.78 (+80%)
Plan difference: Access only in the old plan: rag.Document Index Scan [UQ_Document_Source]; rag.Document Clustered Index Seek [PK_Document] (lookup); rag.Chunk Index Seek [UQ_Chunk_DocSeq]; rag.Chunk Clustered Index Seek [PK_Chunk] (lookup); Access only in the new plan: rag.Chunk Clustered Index Seek [PK_Chunk]; rag.Document Clustered Index Seek [PK_Document]; Join only in the old plan: Merge Join (Inner Join); Nested Loops (Inner Join) ×3; Join only in the new plan: Nested Loops (Inner Join) ×2
Candidate cause: STATS — Statistic rag.Chunk.PK_Chunk was updated 2026-10-07 20:46 (sampled 50.6%); 3 updates in the window; the plans use this statistic.
Candidate cause: STATS — Statistic rag.Chunk.UQ_Chunk_DocSeq was updated 2026-10-07 20:46 (sampled 51.3%); 3 updates in the window; the plans use this statistic.
Candidate cause: STATS — Statistic rag.Chunk._WA_Sys_00000003_49309088 was updated 2026-10-07 20:46 (sampled 50.6%); 3 updates in the window; the plans use this statistic.
Candidate cause: STATS — Statistic rag.Document.PK_Document was updated 2026-10-07 10:27 (sampled 100%); the plans use this statistic.
Candidate cause: STATS — Statistic rag.Document.UQ_Document_Source was updated 2026-10-07 13:59 (sampled 100%); the plans use this statistic.
Candidate cause: STATS — Statistic rag.Document._WA_Sys_0000000A_428392F9 was updated 2026-10-07 20:11 (sampled 100%); 2 updates in the window; the plans use this statistic.
Candidate cause: STATS — Statistic rag.Document._WA_Sys_00000005_428392F9 was updated 2026-10-07 21:00 (sampled 100%); 3 updates in the window; the plans use this statistic.
Candidate cause: STATS — Statistic rag.Document._WA_Sys_0000000E_428392F9 was updated 2026-10-07 13:59 (sampled 100%); the plans use this statistic.
Candidate cause: STATS — Statistic rag.Document._WA_Sys_00000007_428392F9 was updated 2026-10-07 21:02 (sampled 100%); 3 updates in the window; the plans use this statistic.
Counter-evidence: INDEX_UNUSED — rag.Document.IX_Document_Pair was created in the window but the new plan does not use it: candidate ruled out.
Counter-evidence: WORK_PROPORTIONAL — Rows per execution changed 1.8× and duration 1.46×; duration per row 2.0344 → 1.6515 ms: the duration difference comes from the amount of data processed, not the plan.
Confidence rationale: Low: strong counter-evidence (WORK_PROPORTIONAL).
Note: this is a candidate cause, not proof of the root cause.