A read-only SQL Server performance analyzer

Dashboard refreshes an on-demand snapshot of the active connection: CPU, memory, workload, IO, and TempDB pressure, each read against a documented threshold rather than left as a raw number. It is built for the first minute of an investigation — read the current state, find the pressure area, then open the module that explains it.

Nothing is collected while the desktop application is closed. Every figure on the screen was read from the server at the moment you asked for it.

• Server health and CPU pressure
• Memory health and buffer pool condition
• IO latency and TempDB contention
• Workload and active session pressure

Query Statistics ranks the highest-impact queries for the selected database using CPU, duration, and logical read scoring, then lets you open the source and the execution plan from the same row. Query Store regressions surface here too, so a query that got slower after a plan change is visible without hand-writing a DMV query.

Deterministic scoring comes first. AI interpretation is optional, runs against a single query or a selected batch, and produces a tuning report you can export and hand to someone else.

• Top-impact ranking by CPU, duration, and reads
• Execution plan and source inspection
• Query Store regression detection
• Exportable AI tuning reports

Index Advisor is index review and maintenance planning, not a dump of the missing-index DMVs. It classifies each index deterministically, backs that classification with Query Store evidence, and separates what should be kept, maintained, validated, or considered for removal.

Drop safety is treated as a signal in its own right: before an index is put forward for removal, the usage evidence behind that suggestion is shown alongside it. Optional AI interpretation adds context; it never applies a change.

• Deterministic index classification
• Query Store evidence per recommendation
• Drop-safety assessment before removal advice
• Maintenance planning across the database

Blocking Analysis discovers the full blocking chain and ranks head blockers, so the session you act on is the actual root of the contention rather than the loudest victim. Session detail, timeline history captured by the application, and kill-safety signals sit in the same view.

When the incident is over the evidence exports as a structured report — which is what matters when the question the next morning is what happened and who was affected, not simply which session was killed.

• Blocking chain discovery and head-blocker ranking
• Live session impact and kill-safety signals
• Application-captured incident timeline
• Exportable incident evidence

Wait Statistics combines cumulative wait data with the sessions that are waiting right now, so a dominant wait category can be traced to real work instead of read as an abstract percentage. PAGEIOLATCH, CXPACKET, and SOS_SCHEDULER_YIELD are presented with the context needed to tell a symptom from a cause.

Historical trend snapshots let you compare a window against an earlier baseline, and optional query-context correlation connects a wait category back to the queries producing it.

• Cumulative wait categories with active waiters
• Baseline comparison across trend windows
• Blocking-chain correlation
• Query-plan context for a wait type

Security Audit collects server-level and current-database security signals and turns them into structured findings with severity counts, a login inventory, patch status, and an overall maturity level and score.

The whole review is read-only: it reports exposure, it does not remediate it. Findings export as compliance-ready evidence, which is usually what an auditor or a remediation plan actually needs.

• Risky permissions and login inventory
• Server and database configuration exposure
• Patch status and severity breakdown
• Maturity score with exportable findings

Scheduled Jobs is an interactive review surface for SQL Server Agent: job status, recent runs, step-level failures, failure groupings that make a repeating problem obvious, and Database Mail issues that would otherwise quietly hide a broken notification path.

Jobs are never executed and SQL Agent configuration is never modified. You inspect the evidence and decide what to do with it.

• Job outcomes and recent run history
• Step-level failure evidence
• Alert-oriented failure grouping
• Database Mail delivery problems

Object Explorer is the read-oriented browsing surface for the selected database. Locate a procedure, view, or table, inspect its definition, review cached execution statistics, and trace dependencies before anything is changed.

AI Tune can be launched from an object once the surrounding evidence is already in front of you, so interpretation follows inspection instead of standing in for it.

• Procedures, views, tables, and definitions
• Cached execution statistics per object
• Dependency tracing
• AI Tune from the object in context