Docs/Modules/Object Explorer

Object Explorer

Browse SQL Server procedures, views, tables, dependencies, cached execution stats, and AI Tune reports from one object-level investigation workflow.

Overview

Object Explorer is the main browsing and inspection surface for SQL Server objects in the currently selected database. It is built for read-oriented investigation so you can locate an object, inspect its definition, review lightweight statistics, trace dependencies, and launch AI Tune from one workflow.

This page works especially well as the bridge between broad triage and focused tuning. Start here when you know the object you want to inspect, or jump here from Query Statistics or Index Advisor when you need source, metadata, and dependency context around a specific object.

What You Can Do
  • Browse user objects by database and object type.
  • Search by object name with multi-term matching.
  • Inspect source code or generated table definitions.
  • Review cached execution or table-level metadata.
  • Trace dependencies and launch AI-assisted analysis.
Supported Objects
  • Stored Procedures
  • Views
  • Functions
  • Triggers
  • Tables
Screen 1

Source Code and Object Navigation

The left panel is the navigation surface for object discovery, while the Source Code tab on the right is the first inspection layer. This is the fastest way to open a procedure, view, function, trigger, or generated table definition without leaving the product.

Left Panel Controls
Database Filter
  • Loads accessible online databases and excludes system-only cases like tempdb and model.
  • Attempts to preselect the active database from the current connection.
  • Changing the selection refreshes the object list for that database.
Object Type Filter

Narrow the list to All Objects, Stored Procedures, Views, Triggers, Functions, or Tables.

The current implementation maps these options to SQL Server object type codes such as P, V, FN, TR, and U.

Search Box
  • Filters the currently loaded list client-side.
  • Splits the search text into lowercase terms.
  • Keeps only items that match all entered terms.
Object List
  • Displays objects in schema.object_name format with type icons.
  • Stores schema, name, and SQL Server type metadata for downstream tabs.
  • Acts as the main entry point for inspection and analysis.
Screen 2

Statistics Tab for Cached Execution Evidence

For procedures, views, functions, and triggers, the Statistics tab surfaces cached-plan execution evidence such as count, CPU, duration, reads, and last execution. This is a quick triage surface rather than a full historical performance ledger.

Screen 3

Relations Tab for Dependency Tracing

The Relations tab shows both forward and reverse dependencies when SQL Server metadata can resolve them. This is useful for impact tracing, review navigation, and understanding whether a procedure, view, or table sits inside a wider dependency chain.

Detail Tabs
Source Code
  • Loads programmable object definitions from SQL module metadata.
  • Generates a best-effort CREATE TABLE script for tables.
  • Falls back to a read-only message when source is not available.
Statistics
  • Shows cached execution metrics for procedures, views, functions, and triggers.
  • Shows row count, size, columns, indexes, and recent read/write metadata for tables.
  • Cached-plan metrics are useful for quick triage but are not full historical totals.
Relations
  • Shows both Depends On and Used By lists.
  • Supports double-click navigation into related objects.
  • Dependency results are best-effort and may miss dynamic SQL references.
AI Tune
  • Builds context from source, metadata, execution stats, relations, Query Store, and plan signals.
  • Supports comprehensive analysis plus optional force refresh behavior.
  • Provides progress, logs, confidence indicators, and export actions after analysis.

The detail workflow is most effective when combined with Query Statistics for query-level evidence and Index Advisor when the inspected source points toward indexing changes.

Screen 4

AI Tune Analysis Options

AI Tune starts from the selected object and opens an analysis options dialog before execution. This is where the user decides whether to bypass cache and run a fresh object analysis, using all collected evidence that the module can assemble.

Screen 5

AI Tune Results and Confidence Feedback

After analysis, the AI Tune tab becomes an evidence-backed review surface with an executive summary, runtime evidence notes, progress logs, and a confidence badge. The result can then be exported, copied, or re-run for deeper validation.

AI Tune Workflow
Before analysis
  • Select an object from the left list.
  • Open the AI Tune tab.
  • Review the selected object label and collection status.
  • Start analysis after the module prepares the object context.
After analysis
  • Review progress logs, status, result text, and the confidence badge.
  • Re-run the analysis when you need fresher or broader evidence.
  • Export reports as HTML, Markdown, or Text.
  • Save the LLM request payload for audit or debugging when needed.
Example AI Report Downloads

The examples below show how Object Explorer AI Tune output can be exported as standalone HTML reports for review, tuning discussions, or audit trails.

AI Report 1

Stored procedure AI report focused on parameter sensitivity, runtime evidence limits, and tuning sequence.

AI Report 2

Example report for non-SARGable query patterns and source-level remediation guidance.

AI Report 3

Example report that combines object-level source review with filtered-index candidate discussion.

Typical Workflow
  1. Select the target database.
  2. Filter by object type and use search to find the object.
  3. Inspect definition or generated table script in Source Code.
  4. Review cached execution data or table metadata in Statistics.
  5. Check dependencies in Relations.
  6. Open AI Tune for deeper analysis and save the result if needed.
Interpretation Notes and Safety
  • No source definition does not always mean the object is unsupported; tables use generated scripts instead.
  • No execution data in Statistics does not automatically mean the object is unused.
  • Relations can be incomplete when dependency metadata cannot resolve dynamic SQL.
  • AI Tune quality depends on available source, plan, Query Store, and dependency context.
  • The module is read-only and intended for inspection, analysis, and evidence collection.