SQL Performance Visualizer

A desktop tool for SQL Server database maintenance. Connect to any SQL Server instance and drill into databases, tables, columns, indexes, and statistics — down to index fragmentation and the statistics histogram — so you can make informed decisions about index rebuilds, reorganizes, and statistics updates, and trigger them directly from the app. An AI tab surfaces the top resource-consuming queries from Query Store and can capture a query's actual execution plan and have it analyzed by Claude, producing a saved markdown report.

Features

• Connect to any SQL Server using Windows Authentication (default) or SQL Authentication
• Red/green indicator shows connection status at a glance
• Browse databases — system and DW* databases filtered out automatically
• Tables tab — tables with row count, data size (MB), and index size (MB); columns panel alongside with data type and nullability
• Indexes tab — every index on the selected table, with page count, size (MB), and live fragmentation %
  • Rebuild / Reorganize buttons on each row run ALTER INDEX ... REBUILD / REORGANIZE directly, with a small red dot shown while that index is being worked on
• Statistics tab — every statistic on the table (primary key first, then user/index stats, then auto-created _WA_Sys_* stats last), with a detail panel showing:
  • the equivalent DBCC SHOW_STATISTICS command
  • the stat header (rows, rows sampled, steps, density, average key length, filter expression, ...)
  • the density vector and the full histogram
  • an Update Stats button to run UPDATE STATISTICS on just that statistic, with a choice of SAMPLE x PERCENT or FULLSCAN — a red dot marks the row that was just updated
• Column highlight — click a statistics row to highlight the columns it covers, back on the Tables tab
• AI tab — top 25 resource-consuming queries from Query Store, ranked by a selectable metric (total duration / total CPU / logical reads / execution count), with executions, avg duration/CPU (ms), reads, and last execution time per plan
  • Execute + Analyze re-runs the selected query with SET STATISTICS XML ON (after an explicit confirmation popup — the query really executes), captures the actual execution plan, and saves it to Documents\SQLPerformanceVisualizer\plans\<database>\query_<id>_<timestamp>.sqlplan (the Query Store estimated plan is saved alongside as .estimated.sqlplan)
  • The plan is then analyzed headlessly by the Claude Code CLI using Erik Darling's query-plan-analysis skill; the report is saved as a matching .md file and rendered with markdown formatting in the app
  • Step-by-step progress (1/3 → 3/3), a live elapsed clock, and red (running) / green (analyzed) row dots
• European (nl-NL) number formatting for row counts, sizes, pages, and fragmentation percentages

Prerequisites

• .NET 8 SDK
• SQL Server instance (local or remote) — any edition including Express
• For Rebuild/Reorganize and Update Stats: a login with ALTER permission on the table (or db_owner); for the statistics detail panel (DBCC SHOW_STATISTICS), at least SELECT permission on the table (SQL Server 2016 SP1+) or ALTER/db_owner on older versions
• For the AI tab:
  • Query Store enabled on the database: ALTER DATABASE <db> SET QUERY_STORE = ON
  • Claude Code CLI installed and on PATH (claude.exe or the npm claude.cmd shim are both detected)
  • Erik Darling's claude-plugins marketplace installed with the sqlserver-query-plans plugin (provides the query-plan-analysis skill), plus Python for its plan extractor
  • Each analysis is a real (billed) Claude Code invocation and typically takes a few minutes

Build & Run

git clone https://github.com/ronaldgithub/SQLPerformanceVisualizer.git
cd SQLPerformanceVisualizer
dotnet run --project SQLPerformanceVisualizer

Or build first:

dotnet build
dotnet run --project SQLPerformanceVisualizer

Usage

1. Enter the server name or IP in the top bar (defaults to localhost)
2. The dot turns red (disconnected) or green (connected)
3. Click Connect — the Database combo box populates
4. Pick a database — the tables list populates and the first table is selected automatically
5. On the Tables tab, click a table — its columns load and the first column is selected automatically
6. Switch to the Indexes tab to see every index on that table, with size, fragmentation, and Rebuild/Reorganize buttons
7. Switch to the Statistics tab and click a row to see its DBCC SHOW_STATISTICS detail (header, density vector, histogram); click Update Stats to refresh it with a sample percent or full scan
8. Switch to the AI tab, pick a ranking metric, select a query, and click Execute + Analyze — review the SQL in the confirmation popup (the query really runs, including any data modifications), then watch the three steps complete and read the rendered report; the .sqlplan and .md files land in Documents\SQLPerformanceVisualizer\plans\<database>\

UI Layout

┌──────────────────────────────────────────────────────────────────────┐
│  SQL Server: [localhost] ● [Connect]      Database: [▼ combo]        │
├──────────────────────────────────────────────────────────────────────┤
│  [ Tables ]  [ Indexes ]  [ Statistics ]  [ AI ]                      │
├──────────────────────────────────────────────────────────────────────┤
│  Tables tab:                                                          │
│  ┌──────────────────────────────┬───────────────────────────────┐   │
│  │ Table     Rows  Data  Index  │ Column     Type      Null      │   │
│  │ dbo.A    1.234    12      3  │ Id         int       no        │   │
│  │ dbo.B      567     4      1  │ Name       nvarchar  yes       │   │
│  └──────────────────────────────┴───────────────────────────────┘   │
│                                                                        │
│  Indexes tab:                                                         │
│  Index Name  [Rebuild][Reorg]●  Type  Pages  Size  Frag%  Unique  PK  │
│                                              Key Columns   Included   │
│                                                                        │
│  Statistics tab:                                                      │
│  ┌──────────────┬───────────────────────────────────────────────┐   │
│  │ ● Stat list  │ Statistic Detail              [ Update Stats ] │   │
│  │ (red dot =   │ DBCC SHOW_STATISTICS('schema.table','stat');   │   │
│  │  just        │ Stat Header / Density Vector / Histogram       │   │
│  │  updated)    │                                                 │   │
│  └──────────────┴───────────────────────────────────────────────┘   │
│                                                                        │
│  AI tab:                                                              │
│  ┌──────────────┬───────────────────────────────────────────────┐   │
│  │ Query Store  │ AI Plan Analysis  ▓▓▓ 1:23  [Execute+Analyze] │   │
│  │ top queries  │ SQL of the selected query                      │   │
│  │ [metric ▼]   │ AI Analysis Report (rendered markdown)         │   │
│  │ ● = busy     │  — actual plan captured, saved, analyzed by    │   │
│  │ ● = analyzed │    the Claude Code query-plan-analysis skill   │   │
│  └──────────────┴───────────────────────────────────────────────┘   │
└──────────────────────────────────────────────────────────────────────┘

Tech Stack

• Avalonia UI 12 — cross-platform .NET UI framework
• CommunityToolkit.Mvvm — MVVM source generators ([ObservableProperty], [RelayCommand])
• Microsoft.Data.SqlClient — SQL Server data access
• Microsoft.Extensions.DependencyInjection — DI container
• Claude Code CLI (external, optional) — headless AI analysis of execution plans on the AI tab; markdown reports are rendered by a small built-in viewer, no extra NuGet packages