April 1, 2026 Release

What's New

sp_QuickieStore 6.4

• New @find_high_impact parameter — automated diagnostic that identifies high-impact queries, plan instability, and workload concentration
• Workload concentration summary in @find_high_impact output
• Physical reads clarification in @find_high_impact output
• Start/end dates displayed in local time for @find_high_impact
• Fix: plan/object IDs no longer leak into query_id_list in @find_high_impact
• Expert Mode output for sys.dm_db_tuning_recommendations (thanks @ReeceGoding — issue #483)

sp_QuickieStore @find_high_impact performance & output improvements

• Performance overhaul: ~60s → ~10s on large Query Stores
  • Split 4-way DMV joins into staged temp table steps (interval IDs → runtime stats → plan rollup → query hash)
  • Pre-staged query_ids and plan_ids reused across representative text, time bucketing, identifiers, and plan fetch
  • Static SQL INSERT for scoring/diagnostics, separate OUTER APPLY for plans
  • EXISTS semi-join patterns for interval filtering and plan fetch
  • Clustered PKs on 7 temp tables for better optimizer estimates
  • Removed redundant maintenance filters and query_plan IS NOT NULL check
• query_sql_text now uses processing-instruction XML format (clickable in SSMS)
• top_waits column omitted when Query Store wait stats unavailable (SQL 2016 or waits not enabled)
• @sort_order parameter now serves as tiebreaker when impact_score ties (defaults to cpu_share)
• Representative text query rewritten to GROUP BY integer columns instead of nvarchar(max)

sp_HealthParser 3.4

• Added MAXDOP 1 to all SELECT INTO queries for consistent performance

sp_IndexCleanup 2.4

• Fix: merge bug that lost include columns in subset chains
• Fix: unique constraints now recognized as superset targets in Key Subset detection (#721)
• Performance: stage dm_db_index_usage_stats into temp table (~4 min → ~10s)
• Performance: replace ISNULL(t.schema_id, v.schema_id) schema join with sys.objects in all queries

Repo

• Issue templates converted from markdown to YAML issue forms

Version Bumps

All procedures updated to x.4 / 20260401:

March 1, 2026 Release

What's New

sp_HealthParser 3.3

• New @skip_waits parameter to skip the wait stats section
• New @use_ring_buffer parameter to use ring_buffer target instead of file target for faster collection

sp_PerfCheck 2.3

• Rework and README accuracy fixes

sp_PressureDetector 6.3

• 14 new perfmon counters
• Filter out zero-rate perfmon counters

sp_QuickieStore 6.3

• Expert Mode now always shows plan hashes
• Expert Mode T-SQL for plan forcing, unforcing, hint setting, and hint removal (thanks @ReeceGoding — PR #684)
• Expert Mode output for sys.database_automatic_tuning_configurations (thanks @ReeceGoding — issue #661)

Version Bumps

All procedures updated to x.3 / 20260301:

What's Changed

• Add CI pipeline for multi-version SQL Server testing by @erikdarlingdata in #665
• Add @skip_waits parameter to sp_HealthParser by @erikdarlingdata in #666
• Add @skip_waits parameter to sp_HealthParser by @erikdarlingdata in #667
• Add @use_ring_buffer parameter to sp_HealthParser (v3.4) by @erikdarlingdata in #668
• Release sp_HealthParser 3.4 — @use_ring_buffer parameter by @erikdarlingdata in #669
• Fix build automation + add ring buffer target by @erikdarlingdata in #671
• Merge dev: build automation fix + ring buffer target by @erikdarlingdata in #672
• Switch build automation to PR-based workflow by @erikdarlingdata in #673
• Automation: Format and Build SQL File by @erikdarlingdata in #674
• Switch build automation to GitHub App token by @erikdarlingdata in #676
• Release: sp_PerfCheck v2.3 + README accuracy fixes by @erikdarlingdata in #677
• Add 14 perfmon counters to sp_PressureDetector by @erikdarlingdata in #683
• sp_QuickieStore: tweak feature/661-automatic-tuning so we respect our filters by @ReeceGoding in #681
• sp_QuickieStore: never hide plan hashes in Expert Mode. by @ReeceGoding in #680
• sp_QuickieStore: added expert mode T-SQL for plan forcing and hinting by @ReeceGoding in #684
• March 1 2026 release by @erikdarlingdata in #685
• sp_QuickieStore: add sys.database_automatic_tuning_configurations support by @erikdarlingdata in #686
• March 1 2026 release (v2) by @erikdarlingdata in #687

Full Changelog: Updates_20260215...Updates_20260301

🎉 New Stored Procedure: sp_QueryStoreCleanup

Query Store collects a lot of noise — system DMV queries, index maintenance, statistics updates, DBCC commands — all cluttering up your Query Store and making it harder to find the queries you actually care about.

sp_QueryStoreCleanup identifies and removes duplicate and noisy queries from Query Store in any database on your server.

🔍 Text pattern matching to find system and maintenance noise
🧹 Hash-based deduplication (query_hash, plan_hash, or both)
🛡️ Queries with forced plans are always protected
📊 Report-only mode to preview before removing
⏳ Optional age filter to only remove stale queries

/* Preview what would be removed */
EXEC dbo.sp_QueryStoreCleanup
    @database_name = 'YourDatabase',
    @report_only = 1;

/* Remove the noise */
EXEC dbo.sp_QueryStoreCleanup
    @database_name = 'YourDatabase';

Full documentation: sp_QueryStoreCleanup README

🐛 32 Bug Fixes Across All 10 Stored Procedures

Two rounds of deep code review identified and fixed 32 logical bugs. Every fix was validated against SQL Server 2016, 2017, 2019, 2022, and 2025 — 50/50 installs and 50/50 executions, zero regressions.

sp_QuickieStore (5 fixes)

• Fixed cursor skipping databases when @get_all_databases = 1 (missing FETCH NEXT/CONTINUE)
• Fixed compile memory calculation inflated 8x (_kb columns already in KB, not pages)
• Fixed log bytes calculation deflated ~95x (wrong divisor: 100000000 → 1048576)
• Removed duplicate TRUNCATE statements
• Added @@DATEFIRST warning when @workdays = 1 with unsupported regional settings

sp_PressureDetector (4 fixes)

• Fixed missing SET LOCK_TIMEOUT -1 in @cool_new_columns = 1 branch
• Fixed DATEDIFF divide-by-zero in perfmon total_per_second
• Fixed sampled avg_ms_per_wait using incorrect averaging (now uses proper delta calculation)
• Fixed @prefix NULL on Azure SQL causing query failures

sp_IndexCleanup (5 fixes)

• Fixed LEN() trailing space trim causing off-by-one in comma-separated list truncation (switched to DATALENGTH)
• Fixed t.max_length = -1 never matching for (max) columns (should be c.max_length)
• Plus 3 additional fixes from prior review

sp_HumanEvents (4 fixes)

• Fixed memory filter, CATCH cleanup, QUOTENAME schema/table handling

sp_PerfCheck (3 fixes)

• Fixed NULL @processors causing bad detail strings
• Fixed TOKENANDPERMUSERSTORE priority gap (BETWEEN 3 AND 5 → BETWEEN 2 AND 5)
• Wrapped DMV queries in VIEW SERVER STATE guard

sp_QueryReproBuilder (2 fixes)

• Fixed version gates from @new (SQL 2019+) to @sql_2017 for columns available since SQL 2017

sp_HumanEventsBlockViewer (2 fixes)

• Fixed table mode XML handling and DATEADD overflow guard

sp_QueryStoreCleanup (1 fix)

• Fixed COUNT_BIG(*) false positives → COUNT_BIG(DISTINCT) (eliminated 39% false positive rate)

sp_HealthParser (1 fix)

• Fixed XPath inputbuf extraction

sp_LogHunter (1 fix)

• Added archive > 0 guard

🔧 Additional Improvements

Round 2 Fixes (5 additional)

• sp_HealthParser: Added missing N'' unicode prefix on dynamic SQL string
• sp_HumanEventsBlockViewer: Fixed debug label mismatch (#blocking_sh → #blocking_xml_sh)
• sp_PerfCheck: GETDATE() → SYSDATETIME() for uptime calculation consistency
• sp_PressureDetector: Renamed misleading physical_reads column to logical_reads (source was der.logical_reads)
• sp_PressureDetector: Fixed SUBSTRING start position off-by-one in debug PRINT

sp_PressureDetector

• Filter out zero-rate perfmon counters from output

Tested on SQL Server 2016, 2017, 2019, 2022, and 2025 ✅

Bug Fixes

This release includes bug fixes across multiple stored procedures discovered during code review.

sp_IndexCleanup (2.2.5)

• Fixed rule prioritization for Key Duplicates of MAKE UNIQUE indexes (Rule 7.6)
• Fixed MERGE scripts for non-compressible indexes (changed INNER JOIN to LEFT JOIN)
• Added missing semicolons to generated scripts

sp_HumanEvents (7.2.5)

• Fixed duplicate column name blocked_process_report in blocking output (renamed XML data column to blocked_process_report_xml)

sp_HumanEventsBlockViewer (5.2.5)

• Version bump for consistency

sp_LogHunter (3.2.5)

• Fixed RAISERROR format specifier mismatch (%i changed to %I64d for bigint variable)

sp_PerfCheck (2.2.5)

• Fixed file growth calculation formula
• Fixed COUNT_BIG to integer assignment mismatch
• Fixed WHERE clause logic error
• Added missing AS keyword in column alias

sp_PressureDetector (6.2.5)

• Fixed column alias mismatch in UNION (total_avg_stall vs total_avg_stall_ms)
• Fixed bit variable compared to string 'true' instead of 1

sp_QueryReproBuilder (1.2.5)

• Fixed temp table #procedure_object_ids used before creation when wildcard procedure name used
• Fixed SUBSTRING error in parameter extraction when query plans lack ParameterList XML
• Added missing semicolon after END CATCH

sp_HealthParser (3.2.5)

• Added table logging for blocking/deadlocks
• Fixed data types
• Added blocked process attributes

sp_QuickieStore (6.2.5)

• Added total sort options

• Added @target_output parameter to sp_HumanEvents, if you want to use a file instead of the ring buffer.
• Removed the geometry stuff from sp_QuickieStore. I haven't quite figured out how to make it work yet. Also removed some vestigial checks that are no longer required in Azure.
• Tuned WhatsUpLocks to run less terribly when there are a lot of locks
• Added compression type to WhatsUpIndexes
• Added new 2025 memory pressure stuff to sp_PressureDetector and sp_PerfCheck
• Did extensive work on sp_PerfCheck to fix a lot of stuff I was unhappy with. Difficult to list it all.
• Did extensive work on sp_IndexCleanup to make it more reliable and less complicated.

You'll note that the version dates for these is January 15th. While I am releasing this now (early), I probably won't be back looking at this stuff until January 5th or so. Hopefully you won't either!

What's Changed

• burg by @erikdarlingdata in #636
• Dev by @erikdarlingdata in #642
• sp_QuickieStore: Correct window function in lasts by @ReeceGoding in #644
• Dev by @erikdarlingdata in #647

Full Changelog: Updates_20250901...Updates_20251114

This release has a ton of tiny fixes and tweaks that I've been making to complement the training stuff I've been working on.

It also introduces a new stored procedure called sp_QueryReproBuilder, that uses Query Store data to build easy-to-execute parameterized repro queries. This is a bit of an alpha release, so please be gentle when reporting issues.

What's Changed

• Update sp_HumanEventsBlockViewer.sql by @erikdarlingdata in #622
• Dev by @erikdarlingdata in #626
• Fix Incorrect syntax near NULL when using @include_query_hash_totals = 1 by @Kysluss in #629
• Made sp_HumanEventsBlockViewer able to use the ring buffer for system_health. by @ReeceGoding in #632
• Dev by @erikdarlingdata in #635

New Contributors

• @Kysluss made their first contribution in #629

Full Changelog: Updates_20250601...Updates_20250901

What's Changed

• sp_QuickieStore: Appended _ms where it was missing for avg_cpu_time by @ReeceGoding in #600
• sp_QuickieStore: Moved validation of @sort_order to be much earlier on. by @ReeceGoding in #602
• sp_QuickieStore: Moved special sorting columns for special sort order values to be jus… by @ReeceGoding in #604
• sp_QuickieStore: Made plan hash, query hash, or sql handle show when the parameter for filtering them out is passed in by @ReeceGoding in #608
• sp_QuickieStore: Documents that @format_output = 1 removes most decimals. by @ReeceGoding in #606
• Adding support for indexed views in addition to indexes on tables. by @FirstCall42 in #610
• sp_QuickieStore: Move regression validation to just after sort order by @ReeceGoding in #612
• Made sp_HumanEventsBlockViewer not error out when system_health is used by @ReeceGoding in #618

Full Changelog: Updates_20250501...Updates_20250601

What's Changed

• fix for 00 decimals on avg_read/write_stall_ms and total_avg_stall_ms by @VladDBA in #567
• sp_QuickieStore's Readme.md file - Markdown needs the escape character escaped by @VladDBA in #572
• sp_QuickieStore: Consistent PRIMARY KEY syntax. by @ReeceGoding in #578
• sp_QuickieStore: Made skipping maintenance plans optional by @ReeceGoding in #580
• Expand wait_time_percent_of_uptime to accomodate >1,000% waits by @ChrisMayIVCE in #585
• sp_QuickieStore: Multiple changes to how wait stats are handled by @ReeceGoding in #590
• Made filtering by @execution_count be done with a temp table rather than a variable by @ReeceGoding in #593

New Contributors

• @VladDBA made their first contribution in #567
• @ChrisMayIVCE made their first contribution in #585

Full Changelog: Updates_20240404...Updates_20250501

This release fixes a bunch of stuff that I found in deeper testing across different environments.

It also introduces a new stored procedure called sp_PerfCheck, which does a targeted set of checks for performance-related stuff only. No backups, no checkdb, no security, etc. It's built to work with Azure SQLDB too, which is... I guess sort of helpful, given how little you can change while running SQL Server in Safe Mode. I will be adding more specific checks for cloudy things in the future.

For now, check out the readme to learn more.

Also, give it a run. Let me know what you think. I care about your feelings.

What's Changed

• sp_QuickieStore: @utc_offset_string nvarchar(max) changed to nvarchar(6) by @ReeceGoding in #537
• Compression eligibility changes by @FirstCall42 in #549

New Contributors

• @FirstCall42 made their first contribution in #549

Full Changelog: Updates_20250401...Updates_20240404