Conversation
…s table and its columns to 'MS_Description'. This will allow them to be recognized by SSMS and other property-aware third party tools.
Remove where 1 = 1. Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Remove semicolon even though I think Copilot is drunk here Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Remove extra semicolon Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
…ed-Properties Rename the 'Description' extended properties to 'MS_Description'
Ranks all queries (procs and ad-hoc) by how frequently they appear in deadlocks, with victim vs. survivor breakdown. Groups by sql_handle + statement offsets from the execution stack, deduplicates to one frame per participant to avoid double-counting from nested calls. Shows each query's percentage of total deadlocks with a 10% significance cutoff. Closes #3899 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…lock-query sp_BlitzLock: add top deadlock query finding (check 17)
Adding info about how to enable last actual plans.
20260413 readme info about last actual plans
Agent-Logs-Url: https://github.com/BrentOzarULTD/SQL-Server-First-Responder-Kit/sessions/688c8d89-abbf-444f-93d9-b5c9c2e28fb0 Co-authored-by: BrentOzar <245462+BrentOzar@users.noreply.github.com>
Agent-Logs-Url: https://github.com/BrentOzarULTD/SQL-Server-First-Responder-Kit/sessions/4b6609d6-96ec-4fbe-a023-62e8fc6a35e9 Co-authored-by: BrentOzar <245462+BrentOzar@users.noreply.github.com>
…gh-low-configs sp_Blitz: exclude SQL Server version-part configs from Non-Default Server Config findings
Agent-Logs-Url: https://github.com/BrentOzarULTD/SQL-Server-First-Responder-Kit/sessions/38caff71-fd4e-4c35-843b-7cf1fe8b131c Co-authored-by: BrentOzar <245462+BrentOzar@users.noreply.github.com>
Agent-Logs-Url: https://github.com/BrentOzarULTD/SQL-Server-First-Responder-Kit/sessions/04d7c124-a218-4227-bf4a-3a8923ec2245 Co-authored-by: BrentOzar <245462+BrentOzar@users.noreply.github.com>
Agent-Logs-Url: https://github.com/BrentOzarULTD/SQL-Server-First-Responder-Kit/sessions/5c4abca1-268f-4bf9-928e-fe6af94f899d Co-authored-by: BrentOzar <245462+BrentOzar@users.noreply.github.com>
Agent-Logs-Url: https://github.com/BrentOzarULTD/SQL-Server-First-Responder-Kit/sessions/a0acda18-80c1-4d31-a268-7222355e50a9 Co-authored-by: BrentOzar <245462+BrentOzar@users.noreply.github.com>
Agent-Logs-Url: https://github.com/BrentOzarULTD/SQL-Server-First-Responder-Kit/sessions/7f54f118-755b-4c8e-9e39-ce219d78cf03 Co-authored-by: BrentOzar <245462+BrentOzar@users.noreply.github.com>
Agent-Logs-Url: https://github.com/BrentOzarULTD/SQL-Server-First-Responder-Kit/sessions/033fb078-62be-4d8e-b279-d0187c5c2d9e Co-authored-by: BrentOzar <245462+BrentOzar@users.noreply.github.com>
…-component-warnings sp_Blitz: stop reporting deprecated FRK components as “Component Outdated”
Agent-Logs-Url: https://github.com/BrentOzarULTD/SQL-Server-First-Responder-Kit/sessions/7a1c9181-8160-49d9-bb4b-80459e877cd6 Co-authored-by: BrentOzar <245462+BrentOzar@users.noreply.github.com>
…a-file-thresholds sp_Blitz: replace fixed TempDB file-count warning with CPU-relative threshold
Agent-Logs-Url: https://github.com/BrentOzarULTD/SQL-Server-First-Responder-Kit/sessions/e67e721b-4247-4ead-af5c-937f8ddbcdcb Co-authored-by: BrentOzar <245462+BrentOzar@users.noreply.github.com>
…t-server-config-details sp_Blitz: shorten “Non-Default Server Config” details to start with “Default value:”
…indings New @SkipExecutionPlans bit parameter (default 0) lets callers bypass the execution plans result set and jump straight to the findings rollup. When set to 1, the client-output path skips building #available_plans / #dm_exec_query_stats and the sys.dm_exec_text_query_plan calls, and skips the execution plans result set, returning only the deadlock results and the findings rollup. Default behavior is unchanged. Also documented in @Help and surfaced in the @debug parameter dump. Fixes #4013 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
sp_BlitzLock: add @SkipExecutionPlans to skip plans and return only findings (#4013)
Code: - Remove unused @nl variable. - Remove unused @ProductMajor variable (no version-dependent branches use it). - Quote and normalize @LinkedServer before injection into the 4-part EXEC. Strips outer brackets, re-wraps with QUOTENAME so names with periods, spaces, or other special characters produce a valid identifier and the dynamic SQL can't be tampered with via a maliciously-named linked server. - Harden the linked-server CallStack parser. Trim trailing CR/LF/TAB/space that some clients append, then only strip the closing "';" if it's actually present. If the suffix doesn't match, leave @RemoteXml NULL so the existing RemoteEmpty error path fires instead of silently mangling the payload. - Clarify SpillCount comment: count() returns 0 (not NULL) so the row is always emitted; 0 vs 0 matches as "Same" and gets suppressed; non-zero deltas surface via the exact-match carve-out in the diff filter. README: - Fix hex literal examples for @QueryPlanHash from 7-byte (14 hex digits) to 8-byte (16 hex digits) so the samples run without padding surprises. - Drop "parallel index ops" from the sp_configure allowlist list - it's not a real sp_configure option and isn't in #ConfigAllowlist. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
New sp_BlitzPlanCompare
Keeps it out of Install-All-Scripts.sql and Install-Azure.sql, which build their merged bundles from sp_*.sql at the repo root only (the Get-ChildItem call in 'Build Installation Scripts.ps1' is non-recursive, so anything under OptionalScripts/ is skipped automatically). Users who want sp_BlitzPlanCompare install OptionalScripts/sp_BlitzPlanCompare.sql directly. README now spells that out at the top of the proc's section. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds an Installation callout at the top of the sp_BlitzPlanCompare section explaining the proc lives under OptionalScripts/ and is not part of Install-All-Scripts.sql / Install-Azure.sql, so users know they need to install it explicitly. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Azure Data Studio is retired; SSMS is the only client we should be pointing users at in fresh docs. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
…optional Move sp_BlitzPlanCompare.sql to OptionalScripts/
…kieStore Fixes #4010. Check 8's "More Info - Query" finding for stored-procedure deadlocks used to emit an EXECUTE sp_BlitzQueryStore call - that proc is deprecated and no longer maintained. Per the issue discussion the right replacement is Erik Darling's sp_QuickieStore (https://erikdarling.com/sp_quickiestore/), which has the same shape: a Query Store explorer keyed by database + procedure name. Mapping: sp_BlitzQueryStore @DatabaseName -> sp_QuickieStore @database_name sp_BlitzQueryStore @StoredProcName -> sp_QuickieStore @procedure_name The Query Store availability guard (@ProductVersionMajor >= 13 OR @Azure = 1) stays - sp_QuickieStore also requires Query Store, which is SQL 2016+. Also updated: - the section comment at line 2809 - the debug RAISERROR message at line 2894 so future readers don't see stale "BlitzQueryStore" references. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Copilot review on PR #4017 flagged that ds.database_name comes from PARSENAME(ds.proc_name, 3), which returns NULL whenever the proc_name isn't a 3-part identifier. QUOTENAME(NULL, '''') propagates NULL through the rest of the string concat, so the entire finding becomes NULL and the user gets nothing to copy/paste. Switched to dow.database_name (derived from dow.database_id at deadlock-capture time, already used 4 lines up for the output's database_name column). Same data, more reliable source, stays consistent with the rest of the SELECT. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
…erystore sp_BlitzLock: replace deprecated sp_BlitzQueryStore call with sp_QuickieStore
…ile suffix
Replace PATINDEX('%_[0-9][0-9]%', REVERSE(...)) with a CASE expression that
checks the last underscore position. When the underscore is <=7 chars from the
end (indicating a 1-2 digit split index like _1.bak or _10.bak), strip from
there. Otherwise strip only the file extension (.bak/.trn).
This prevents garbled datetime extraction from non-split filenames where digit
pairs inside the datetime portion (e.g. _190000) were incorrectly matched as
split-file suffixes.
Affected: sp_DatabaseRestore.sql, Install-All-Scripts.sql
…ck, revert Install-All-Scripts.sql - Remove stale PATINDEX line in _[0-9][0-9].bak @StoPAt DELETE block that was left alongside the new CASE WHEN line, causing a T-SQL parse error (AND after ;) - Drop the non-striped .bak DELETE block added by mistake: it reintroduces the same broken PATINDEX pattern and hardcodes .bak instead of @FileExtensionBak; not among the 13 locations listed in #4019 - Revert Install-All-Scripts.sql to pre-PR state — it is auto-generated by the build process and will pick up the sp_DatabaseRestore.sql fix after merge
…tion Fix DIFF backup silently skipped when filename has no numeric split-file suffix
Fixes from the Copilot review on #4002: - Skip the timing recalc when NodeId values are not unique across the document. NodeId is only unique per statement, so on multi-statement plans //p:RelOp[@nodeid=...] would blend operators from different statements and corrupt the rewritten timings. Detected via count(distinct-values(//p:RelOp/@nodeid)) vs count(//p:RelOp). Attribute/element stripping and whitespace minification still apply. - NULL-safe SqlHandle/QueryHash matching when writing ai_query_plan back to ##BlitzCacheProcs. Parent procedure/batch rows carry a NULL QueryHash, so the previous equality match silently discarded their minimized text. - Aggregate ActualExecutions (summed across threads) into the synthesized per-operator summary instead of hardcoding 1, preserving the how-many-times-did-this-run signal for nested loop inner sides. - Count failed .modify() steps per plan and report via RAISERROR when @debug >= 1 instead of swallowing them invisibly. - Document why the 500-RelOp recalc cap exists and what plans over the cap still get. - Extend the @reanalyze schema-drift guard to all five ai_* columns, not just ai_query_plan. Verified on Azure SQL DB: full-proc run shows parent rows (NULL QueryHash) now get ai_query_plan populated and multi-statement proc plans keep their original per-thread blocks; standalone harness run of the verbatim minimizer block confirms ActualExecutions="8" (3+5 across threads) on a DOP-2 nested-loop inner side, CPU summed, elapsed maxed. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Per Copilot round-two review on #4002. (SqlHandle, QueryHash) is not unique in ##BlitzCacheProcs when the same query hash has multiple distinct plans, so the minimizer's UPDATE stamped the last-fetched plan's minimized text onto every row sharing the pair. Add PlanHandle to the cursor key, both FETCH statements, and the NULL-safe UPDATE predicate, mirroring the AI-call cursor's matching. Verified on Azure with two rows sharing (SqlHandle, QueryHash) but different PlanHandles and distinguishable plans: each row's ai_query_plan now contains its own plan's operators only. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…-1a291a sp_BlitzCache: minimize query plan XML sent to AI (#3862)
For latest CUs and GDRs, prepping for release.
Updating build numbers
Member
Author
|
@codex review |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 0c1340da82
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| WHERE BackupFile LIKE N'%[_][0-9][0-9].bak' | ||
| WHERE BackupFile LIKE N'%[_][0-9][0-9].' + @FileExtensionBak | ||
| AND BackupFile LIKE N'%' + @Database + N'%' | ||
| AND (REPLACE( RIGHT( REPLACE( BackupFile, RIGHT( BackupFile, CASE WHEN PATINDEX( '%[_]%', REVERSE( BackupFile ) ) <= 7 THEN PATINDEX( '%[_]%', REVERSE( BackupFile ) ) ELSE CHARINDEX( '.', REVERSE( BackupFile ) ) END ), '' ), 18 ), '_', '' ) > @StopAt); |
There was a problem hiding this comment.
This new predicate ends the DELETE statement with a semicolon, leaving the following AND (...) as a standalone statement. The procedure body will fail to parse with a syntax error near AND during ALTER PROCEDURE, regardless of whether @StopAt is used at runtime.
Useful? React with 👍 / 👎.
| r.blocking_session_id <> 0 | ||
| OR blocked.session_id IS NOT NULL | ||
| OR wt.wait_info IS NOT NULL | ||
| OR (s.status = ''sleeping'' AND COALESCE(s.open_transaction_count, r.open_transaction_count, blocked.open_tran, 0) >= 1) |
There was a problem hiding this comment.
With @OnlyProblems = 1 and the default @ShowSleepingSPIDs = 0 path used by sp_BlitzFirst @ExpertMode = 3, sleeping sessions with open transactions are already removed by the earlier @ShowSleepingSPIDs predicate before this new OR can match them. That makes the advertised “sleeping with an open transaction” problem case invisible unless callers also know to enable sleeping SPIDs, so this predicate should either bypass or adjust the earlier sleeping-session filter.
Useful? React with 👍 / 👎.
| IF @DatabaseName IS NULL OR @DatabaseName = @CurrentDbName | ||
| BEGIN | ||
| /* Current-DB resolution. OBJECT_ID parses bare / 2-part / 3-part. */ | ||
| SET @ResolvedProcObjectId = OBJECT_ID(@StoredProcName); |
There was a problem hiding this comment.
When a caller follows the help text and passes [db].[schema].[proc] without also setting @DatabaseName, this resolves the object_id but leaves @DatabaseName NULL, so the later plan-cache predicate filters only on st.objectid. Object IDs are unique only within a database, so a cached plan from another database with the same object_id can be selected or create spurious ambiguity; parse the database from the three-part name or add the dbid predicate for that case.
Useful? React with 👍 / 👎.
| a period, space, or other special character still produce a valid identifier | ||
| and the dynamic SQL can't be tampered with via a maliciously-named linked server. */ | ||
| DECLARE @LinkedServerSafe SYSNAME = REPLACE(REPLACE(@LinkedServer, N'[', N''), N']', N''); | ||
| SET @sql = N'EXEC ' + QUOTENAME(@LinkedServerSafe) + N'.master.dbo.sp_BlitzPlanCompare @QueryPlanHash = @hash;'; |
There was a problem hiding this comment.
In linked-server mode, even when the local lookup was scoped with @DatabaseName, this remote call sends only the local query hash and drops that database scope. If the same query text exists in multiple databases on the remote server, the remote side can return an ambiguity error or compare against a plan from the wrong database; include the requested database scope in the remote invocation.
Useful? React with 👍 / 👎.
This was referenced Jul 2, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
I'm not actually going to merge this yet, but I wanna trigger a Codex code review for the changes in this month's release, since it's pretty wide-ranging.