Merged
Conversation
Shorten update link text
Delete .coderabbit.yaml
- Increase metric label font size from 9 to 12 for readability - Use ForegroundBrush instead of muted SlicerLabelBrush for metric label - Fix metric label showing "Total" for avg metrics (e.g., avg-cpu showed "Total CPU" instead of "Avg CPU") by splitting combined or-patterns - Add per-bucket average properties to QueryStoreTimeSlice so the slicer chart shape reflects the selected metric (total vs avg) instead of always showing totals Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Fix Query Store time slicer metric display
…ait profile with v/% toggle, resizable separator with chevron, ribbon click/double-click interactions
- display simple stacked wait profile per default instead on %stacked
Most Debug.WriteLine statements for wait stats have been removed, leaving only exception logging. The WaitProfile column visibility toggle now uses SortMemberPath instead of a hardcoded index, making the code more robust to column order changes.
Wrapped denominator in NULLIF to avoid division by zero errors when calculating wait_ratio in QueryStoreService.cs. Also made minor formatting and comment improvements for clarity.
Feature/query store waitstats
- Replace tab indentation with spaces in FetchPlanWaitStatsAsync doc comments to match file convention - Rewrite doc comment for clarity (explain why plan-level WTR differs) - Use existing _fetchCts instead of creating untracked CancellationTokenSource in OnWaitStatsCollapsedChanged Follow-up to PR #137. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- tooltips for data and time in RangeTimeSlicer - vertical bands to display days (even with gaps)
- allow to move RangeTimeSlicer period using keyboards arrows left and right
Avalonia.Skia 11.3.12 transitively pulls SkiaSharp.NativeAssets.Linux 2.88.9 (libSkiaSharp.so v88.1), but the managed SkiaSharp 3.119.0 requires native libs in the [119.0, 120.0) range. At publish time the old .so overwrites the correct one, causing a TypeInitializationException on startup. Pin SkiaSharp.NativeAssets.Linux to 3.119.0 so NuGet resolution picks the correct native library. macOS and Win32 were already at 3.119.0. Fixes #139 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Fix wait stats cleanup: tab indentation and CTS leak
Fix SkiaSharp native library mismatch on Linux (#139)
…improvments Feature/query store timerange improvments
sys.query_store_wait_stats only exists on SQL Server 2017+ (major >= 14) and Azure SQL DB. Additionally, wait stats capture can be disabled per database via WAIT_STATS_CAPTURE_MODE = OFF. Changes: - Add SupportsQueryStoreWaitStats property to ServerMetadata (major >= 14 or Azure) - Add IsWaitStatsCaptureEnabledAsync to QueryStoreService — checks wait_stats_capture_mode_desc in sys.database_query_store_options - Gate QueryStoreGridControl: accept supportsWaitStats flag, hide wait stats panel/column/splitter when unsupported, skip all fetch calls - Wire both checks in QuerySessionControl before constructing the grid On SQL 2016 or when capture is OFF: no wait stats queries are issued, the panel is collapsed, and the Wait Profile column is hidden. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Gate Query Store wait stats for SQL 2017+ with capture mode check
Three changes from community feedback (41s → 3s on slow machines): 1. Remove WHERE p.query_plan IS NOT NULL from the ranked CTE — this predicate on nvarchar(max) causes an expensive implicit conversion scan. NULL plans are already handled by the C# reader (skip row). 2. Two-phase approach: materialize TOP N into #top_plans first using only numeric columns, then join to query_text/plan XML tables for just the winners. Avoids dragging expensive nvarchar(max) columns through the entire CTE evaluation. 3. Remove OPTION (LOOP JOIN) — reported 3x faster without it on multiple test environments. Tested on SQL2022 with PerformanceMonitor Query Store data. Closes #143 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Improve Query Store grid query performance (#143)
Multi-phase materialization approach for drastically better performance on large Query Store datasets: Phase 1: Pre-filter matching interval IDs into #intervals (clustered PK). All subsequent phases use EXISTS semi-join against this tiny table instead of re-evaluating the time predicate. Phase 2: Aggregate runtime_stats by plan_id into #plan_stats (clustered PK on plan_id). Uses EXISTS against #intervals — semi-join avoids materializing the full interval join. Phase 3: Rank best plan per query_id, materialize TOP N into #top_plans. Still no nvarchar(max) columns touched. Phase 4: Hydrate only the TOP N winners with text, plan XML, and metadata. Uses TRY_CONVERT for safe plan XML handling. Additional improvements: - Both time filter paths now use rsi.start_time (indexed) instead of rs.last_execution_time (requires scanning all runtime_stats rows) - OPTION(RECOMPILE) on aggregation phases prevents parameter sniffing on date range parameters producing stale plans - Clustered PKs on temp tables enable index seeks in subsequent joins Tested on SQL2022: ~22ms total for all 4 phases (24 intervals, 546 plans). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Rewrite Query Store grid query using sp_QuickieStore patterns (#143)
… plan grid and try improve performance of querystore grid fetch
FetchPlanWaitStatsAsync now accepts an optional planIds parameter. When provided, the method creates a temp table and filters results to only those plan IDs for efficient querying. SQL construction and connection logic were refactored to support this feature.
… cause grid query crash
…fixes-202603 Fix/querystore improvment and fixes 202603
- Fix fire-and-forget: globalWaitTask variable replaced with _ = discard - Remove Debug.WriteLines from wait stats catch blocks, add separate OperationCanceledException handler for clean cancellation - Remove unused System.Diagnostics using - Restore ps. prefix on orderClause (used in ranked CTE where ps is the correct alias for #plan_stats) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Post-merge cleanup for PR #148
Two regressions introduced in the time slicer PR (#114): 1. Search filter (module, query_id, etc.) was built via BuildSearchFilter() but never passed to FetchTopPlansAsync — the filter parameter was dropped when the call site switched to named parameters for startUtc/ endUtc. Add filter: back to the call. 2. Clicking Fetch reloads the slicer via LoadTimeSlicerDataAsync, which calls LoadData without preserving the current selection. LoadData defaults to last 24h when no selection is provided, discarding the user's custom range. Pass the current _slicerStartUtc/_slicerEndUtc through to LoadData's selectionStart/selectionEnd parameters. Fixes #147 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…r-reset Fix search filter not applied and slicer reset on Fetch (#147)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add GitHub Sponsors funding link
…-v1.1 Feature/query store timeslicer v1.1
- Modify WaitStatsProfileControl to default to ribbon chart mode - Remove the grid data button from WaitStatsProfileControl header - Add legend button to WaitStatsProfileControl that opens a color legend popup - Add vertical dashed lines for day boundaries in WaitStatsRibbonControl - Align X-axis ticks/labels to day boundaries (00:00) in WaitStatsRibbonControl - Add horizontal dashed average line with label in WaitStatsRibbonControl - Apply FormatWaitRatio in WaitStatsRibbonControl tooltips - Apply FormatWaitRatio in WaitProfileBarControl tooltips - Apply FormatWaitRatio in WaitStatsProfileControl grid table rows
…er light grey background - switch wait color legend to the right of the global wait header
…-v1.1 Feature/query store wait stats v1.1
…pping - Remove duplicate `using System;` in WaitStatsProfileControl - Clamp avg line label position to prevent clipping at top of chart - Skip day boundary labels when too dense to prevent overlap on wide ranges Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Fix wait stats: duplicate using, label overlap, avg clipping
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
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
2 participants
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.
Merges
devintomainto trigger the v1.3.0 release workflow.What's in this release
🤖 Generated with Claude Code