SQL Server Performance Monitor

Free, open-source monitoring that replaces the tools charging you thousands per server per year. 30+ collectors, real-time alerts, built-in MCP server for AI analysis. Nothing phones home. Your data stays on your server and your machine.

Supported: SQL Server 2016–2025 | Azure SQL Managed Instance | AWS RDS for SQL Server | Azure SQL Database (Lite only)

---

Download

👉 Not sure which edition to pick? Start with Lite. One download, nothing installed on your server, data flowing in under 5 minutes.

	Full Edition	Lite Edition
What it does	Installs a PerformanceMonitor database with 30 T-SQL collectors running via SQL Agent. Separate dashboard app connects to view everything.	Single desktop app that monitors remotely. Stores data locally in DuckDB + Parquet. Nothing touches your server.
Best for	Production 24/7 monitoring, long-term baselining	Quick triage, Azure SQL DB, locked-down servers, consultants, firefighting
Requires	SQL Agent running (see permissions)	VIEW SERVER STATE (see permissions)
Get started	Run the installer, open the dashboard	Download, run, add a server, done

Both editions include real-time alerts (system tray + email + webhooks), charts and graphs, dark and light themes, CSV export, and a built-in MCP server for AI-powered analysis with tools like Claude.

All release binaries are digitally signed via SignPath — no more Windows SmartScreen warnings.

---

What People Are Saying

"You guys make us DBAs look like absolute rockstars. I'm over here getting showered with praise, and all I do is use your scripts and follow your advice."

"replaced SentryOne and had it running in 10 minutes"

"I've had enough time to gather data and converse with Claude on this. It helped a lot to zone in on CPU starvation from the hypervisor on which the VM runs. IT team currently investigating the host configuration."

---

What You Get

🔍 32 specialized T-SQL collectors running on configurable schedules with named presets (Off, Aggressive, Balanced, Low-Impact) — wait stats, query performance, blocking chains, deadlock graphs, memory grants, file I/O, tempdb, perfmon counters, FinOps/capacity, and more. Query text and execution plan collection can be disabled per-collector for sensitive environments. Switch presets with a pair of SQL Agent jobs to get quiet-hours / overnight windows without writing any code.

🚨 Real-time alerts for blocking, deadlocks, and high CPU — system tray notifications, styled HTML emails with full XML attachments, and webhook notifications for external integrations

📊 NOC-style dashboard with green/yellow/red health cards, auto-refresh, configurable time ranges, and dark/light themes

📋 Graphical plan viewer with native ShowPlan rendering, 30-rule PlanAnalyzer, operator-level cost breakdown, and a standalone mode for opening .sqlplan files without a server connection

🤖 Built-in MCP server with 51-63 read-only tools for AI analysis — ask Claude Code or Cursor "what are the top wait types on my server?" and get answers from your actual monitoring data

🧰 Community tools installed automatically — sp_WhoIsActive, sp_BlitzLock, sp_HealthParser, sp_HumanEventsBlockViewer

🔒 Your data never leaves — no telemetry, no cloud dependency, no phoning home. Credentials stored in Windows Credential Manager with OS-level encryption.

---

More Screenshots

Lite Edition — Query Performance

Graphical Plan Viewer

Alert Notifications

MCP Server — AI-Powered Analysis

---

Quick Start — Lite Edition

1. Download and extract PerformanceMonitorLite (requires .NET 8 Desktop Runtime)
2. Run PerformanceMonitorLite.exe
3. Click + Add Server, enter connection details, test, save
4. Double-click the server in the sidebar to connect

Data starts flowing within 1–5 minutes. That's it. No installation on your server, no Agent jobs, no sysadmin required.

Upgrading from zip? Click Import Settings then Import Data in the sidebar and point both at your old Lite folder. Settings imports server connections, alert thresholds, SMTP config, and schedules. Data imports historical DuckDB + Parquet archives. Auto-update users (installed via Setup.exe) get updates automatically — no manual import needed.

Always On AG? Enable ReadOnlyIntent in the connection settings to route Lite's monitoring queries to a readable secondary, keeping the primary clear. Enable MultiSubnetFailover for multi-subnet failover scenarios.

Lite Collectors

24 collectors run on independent, configurable schedules:

Collector	Default	Source
query_snapshots	1 min	sys.dm_exec_requests + sys.dm_exec_sessions
blocked_process_report	1 min	XE ring buffer session
waiting_tasks	1 min	sys.dm_os_waiting_tasks
wait_stats	1 min	sys.dm_os_wait_stats (deltas)
query_stats	1 min	sys.dm_exec_query_stats (deltas)
procedure_stats	1 min	sys.dm_exec_procedure_stats (deltas)
cpu_utilization	1 min	sys.dm_os_ring_buffers scheduler monitor
file_io_stats	1 min	sys.dm_io_virtual_file_stats (deltas)
memory_stats	1 min	sys.dm_os_sys_memory + memory counters
memory_grant_stats	1 min	sys.dm_exec_query_memory_grants
tempdb_stats	1 min	sys.dm_db_file_space_usage
perfmon_stats	1 min	sys.dm_os_performance_counters (deltas)
deadlocks	1 min	system_health Extended Events session
session_stats	1 min	sys.dm_exec_sessions active session tracking
memory_clerks	5 min	sys.dm_os_memory_clerks
memory_pressure_events	5 min	sys.dm_os_ring_buffers RING_BUFFER_RESOURCE_MONITOR
query_store	5 min	Query Store DMVs (per database)
running_jobs	5 min	msdb job history with duration vs avg/p95
database_size_stats	15 min	sys.master_files + FILEPROPERTY + dm_os_volume_stats
server_properties	15 min	SERVERPROPERTY() hardware and licensing metadata
server_config	On connect	sys.configurations
database_config	On connect	sys.databases
database_scoped_config	On connect	Database-scoped configurations
trace_flags	On connect	DBCC TRACESTATUS

Lite Data Storage

All data is stored in %LOCALAPPDATA%\PerformanceMonitorLite\ — separate from the executable, so auto-updates don't affect your data.

• Hot data in DuckDB 1.5.2 — non-blocking checkpoints, free block reuse, stable file size without periodic resets
• Archive to Parquet with ZSTD compression (~10x reduction) — automatic monthly compaction keeps file count low (~75 files vs thousands)
• Retention: 3-month calendar-month rolling window
• Typical size: ~50–200 MB per server per week

Lite Configuration

All configuration lives in the config/ folder:

File	Purpose
servers.json	Server connections (passwords in Windows Credential Manager). Optional Utility Database per server for community procs installed outside master.
settings.json	Retention, MCP server, startup behavior, alert thresholds, SMTP configuration
collection_schedule.json	Per-collector enable/disable and frequency
ignored_wait_types.json	144 benign wait types excluded by default

---

Quick Start — Full Edition

Install

Windows Authentication:

PerformanceMonitorInstaller.exe YourServerName

SQL Authentication:

PerformanceMonitorInstaller.exe YourServerName sa YourPassword

Entra ID (MFA) Authentication:

PerformanceMonitorInstaller.exe YourServerName --entra user@domain.com

Clean reinstall (drops existing database and all collected data):

PerformanceMonitorInstaller.exe YourServerName --reinstall
PerformanceMonitorInstaller.exe YourServerName sa YourPassword --reinstall

Uninstall (removes database, Agent jobs, and XE sessions):

PerformanceMonitorInstaller.exe YourServerName --uninstall
PerformanceMonitorInstaller.exe YourServerName sa YourPassword --uninstall

The installer automatically tests the connection, checks the SQL Server version (2016+ required), executes SQL scripts, downloads community dependencies, creates SQL Agent jobs, and runs initial data collection. You can also install directly from the Dashboard's Add Server dialog.

Air-gapped environments? Place pre-downloaded community scripts (sp_WhoIsActive.sql, DarlingData.sql, Install-All-Scripts.sql) in a community/ directory next to the installer. The installer uses local files when present and falls back to GitHub downloads otherwise.

CLI Installer Options

Option	Description
SERVER	SQL Server instance name (positional, required)
USERNAME PASSWORD	SQL Authentication credentials (positional, optional)
--entra EMAIL	Microsoft Entra ID interactive authentication (MFA)
--reinstall	Drop existing database and perform clean install
--uninstall	Remove database, Agent jobs, and XE sessions
--reset-schedule	Reset collection schedule to recommended defaults
--preserve-jobs	Keep existing SQL Agent job schedules during upgrade
--encrypt=optional|mandatory|strict	Connection encryption level (default: mandatory)
--trust-cert	Trust server certificate without validation (default: require valid cert)
--help	Show usage information and exit

Environment variable: Set PM_SQL_PASSWORD to avoid passing the password on the command line.

Exit Codes

Code	Meaning
0	Success
1	Invalid arguments
2	Connection failed
3	Critical file failed (scripts 01–03)
4	Partial installation (non-critical failures)
5	Version check failed (SQL Server 2014 or earlier)
6	SQL files not found
7	Uninstall failed
8	Upgrade script failed

Post-Installation

1. Ensure SQL Server Agent is running — the collection job executes every minute
2. Verify installation:

SELECT * FROM PerformanceMonitor.config.current_version;

SELECT TOP (20) *
FROM PerformanceMonitor.config.collection_log
ORDER BY collection_time DESC;

3. Launch the Dashboard (Dashboard/ folder — build with dotnet build or use the release package). The Dashboard is a separate WPF application that runs on your workstation and connects to any SQL Server where the PerformanceMonitor database is installed. Add your server, enter credentials, and data appears immediately.

What Gets Installed

• PerformanceMonitor database with collection tables and reporting views
• 32 collector stored procedures for gathering metrics (including SQL Agent job monitoring)
• Configurable collection — query text and execution plan capture can be disabled per-collector via config.collection_schedule (collect_query, collect_plan columns) for sensitive or high-volume environments
• Delta framework for calculating per-second rates from cumulative DMVs
• Community dependencies: sp_WhoIsActive, sp_HealthParser, sp_HumanEventsBlockViewer, sp_BlitzLock
• SQL Agent jobs: Collection (every 1 minute), Data Retention (daily at 2:00 AM), and Hung Job Monitor (collection job watchdog, every 5 minutes)
• Version tracking in config.installation_history

Data Retention

Default: 30 days (configurable per collector via the retention_days column in config.collection_schedule).

Storage estimates: 5–10 GB per week, 20–40 GB per month.

Managed Platform Support

The Full Edition supports Azure SQL Managed Instance and AWS RDS for SQL Server with some limitations:

Feature	On-Premises	Azure SQL MI	AWS RDS
All core collectors	Yes	Yes	Yes
Default trace collectors	Yes	Disabled automatically	Yes
System health XE (file target)	Yes	Disabled automatically	Yes
SQL Trace collectors	Yes	Disabled automatically	Yes
SQL Agent jobs	Yes	Yes	Yes
Running jobs collector	Yes	Yes	Disabled automatically
Blocked process threshold	Auto-configured	Auto-configured	Configure via RDS parameter group
sp_configure	Yes	Yes	Not available

Azure SQL MI: The installer automatically detects Engine Edition 8 and disables 4 collectors that require file system access or SQL Trace (default_trace, trace_management, trace_analysis, system_health). All other collectors work normally.

AWS RDS: The installer automatically detects the rdsadmin database and disables the running_jobs_collector (requires msdb.dbo.syssessions which is restricted on RDS). It also gracefully handles restricted sp_configure and limited msdb permissions. SQL Agent jobs are created and owned by the installing login. The RDS master user is automatically enrolled in SQLAgentUserRole; for other logins, add them to SQLAgentUserRole in msdb before running the installer.

AWS RDS Parameter Group Configuration

sp_configure is not available on AWS RDS for SQL Server. Features that depend on server-level configuration must be set through AWS RDS Parameter Groups instead.

Blocked process threshold — Required for blocked process report collection. Without this, blocked process reports will not fire on RDS.

1. Open the AWS RDS Console and navigate to Parameter groups

2. Create a new parameter group (or modify the one attached to your instance):

   • Family: sqlserver-ee-16.0 (or your edition/version)
   • Type: DB Parameter Group

3. Search for blocked process threshold (s) and set it to 5 (seconds)

4. Apply the parameter group to your RDS instance (may require a reboot if the parameter is static)

5. Verify it took effect:

   SELECT
       c.name,
       c.value_in_use
   FROM sys.configurations AS c
   WHERE c.name = N'blocked process threshold (s)';
   

Deadlocks — No parameter group configuration is required. The SQL Server deadlock monitor runs automatically on all platforms, and the xml_deadlock_report Extended Event fires without any threshold setting.

Azure SQL Database — The blocked process threshold is fixed at 20 seconds and cannot be changed. The blocked_process_report event fires automatically when blocking exceeds this duration.

---

Edition Comparison

Capability	Full	Lite
Target server installation	Required	None
SQL Server Agent	Required	Not needed
Azure SQL Managed Instance	Supported	Supported
AWS RDS for SQL Server	Supported	Supported
Azure SQL Database	Not supported	Supported
Multi-server from one seat	Per-server install	Built-in
Collectors	32	24
Agent job monitoring	Duration vs historical avg/p95	Duration vs historical avg/p95
Data storage	SQL Server (on target)	DuckDB + Parquet (local)
Execution plans	Collected and stored (can be disabled per-collector)	Download on demand
Graphical plan viewer	Built-in with 30-rule PlanAnalyzer	Built-in with 30-rule PlanAnalyzer
Standalone plan viewer	Open/paste/drag .sqlplan files	Open/paste/drag .sqlplan files
Community tools (sp_WhoIsActive, sp_BlitzLock)	Installed automatically	Not needed
Alerts (tray + email + webhooks)	Blocking, deadlocks, CPU	Blocking, deadlocks, CPU
Dashboard	Separate app	Built-in
Themes	Dark and light	Dark and light
Portability	Server-bound	Single executable
MCP server (LLM integration)	Built into Dashboard (63 tools)	Built-in (52 tools)

---

Dashboard Tabs

Full Edition Dashboard

Tab	Contents
Overview	Resource overview, daily summary, critical issues, server config changes, database config changes, trace flag changes, collection health
Performance	Performance trends, expensive queries, active queries, query stats, procedure stats, Query Store, Query Store regressions, query trace patterns, query heatmap
Resource Metrics	Server trends, wait stats, TempDB, file I/O latency, perfmon counters, default trace events, trace analysis, session stats, latch stats, spinlock stats
Memory	Memory overview, grants, clerks, plan cache, memory pressure events
Locking	Blocking chains, deadlocks, blocking/deadlock trends
System Events	Corruption events, contention, errors, I/O issues, scheduler issues, memory conditions

Plus a NOC-style landing page with server health cards (green/yellow/red severity indicators).

Lite Edition Dashboard

Tab	Contents
Overview	2x2 resource chart grid (CPU, Memory, Wait Stats, TempDB) with drill-down
Active Queries	Running queries with session details, wait types, blocking, DOP, memory grants
Wait Stats	Filterable wait statistics chart with delta calculations
CPU	SQL Server CPU vs Other Processes over time
Memory	Physical memory overview, SQL Server memory trend, memory clerk breakdown, memory pressure events
Queries	Performance trends, top queries and procedures by duration, Query Store integration, query heatmap
File I/O	Read/write I/O trends per database file
TempDB	Space usage breakdown and TempDB file I/O
Blocking	Blocking/deadlock trends, blocked process reports, deadlock history
Perfmon	Selectable SQL Server performance counters over time
Configuration	Server configuration, database configuration, scoped configuration, trace flags
FinOps	Utilization & provisioning analysis, database resource breakdown, storage growth (7d/30d), idle database detection, index analysis via sp_IndexCleanup, application connections, server inventory, cost optimization recommendations (enterprise feature audit, CPU/memory right-sizing, compression savings, dormant databases, dev/test detection), column-level filtering on all grids

Both editions feature auto-refresh, configurable time ranges, chart drill-down to Active Queries, right-click CSV export, system tray integration, dark and light themes, and timezone display options (server time, local time, or UTC).

---

Alerts & Notifications

Both editions include a real-time alert engine that monitors for performance issues and sends notifications via system tray balloons and email.

Alert Types

Metric	Default Threshold	Description
Blocking	30 seconds (Full), 5 seconds (Lite)	Fires when the longest blocked session exceeds the threshold
Deadlocks	1	Fires when new deadlocks are detected since the last check
Poison waits	100 ms avg	Fires when any poison wait type exceeds the average-ms-per-wait threshold
Long-running queries	5 minutes	Fires when any query exceeds the elapsed-time threshold
TempDB space	80%	Fires when TempDB usage exceeds the percentage threshold
Long-running agent jobs	3× average	Fires when a job's current duration exceeds a multiple of its historical average
High CPU	90% (Full), 80% (Lite)	Fires when total CPU (SQL + other) exceeds the threshold
Server unreachable	N/A	Fires when a monitored server goes offline or comes back online (tray + email)

All thresholds are configurable in Settings.

Poison wait types monitored: THREADPOOL (worker thread exhaustion), RESOURCE_SEMAPHORE (memory grant pressure), and RESOURCE_SEMAPHORE_QUERY_COMPILE (compilation memory pressure). These waits indicate severe resource starvation and should never occur under normal operation.

Notification Channels

• System tray — balloon notifications with a configurable per-metric cooldown (default: 5 minutes)
• Email (SMTP) — styled HTML emails with a configurable per-metric cooldown (default: 15 minutes), plus configurable SMTP settings (server, port, SSL, authentication, recipients)
• Webhook — HTTP POST to a configurable endpoint for integration with external alerting systems (Slack, Teams, PagerDuty, etc.)

All cooldown periods are independently configurable in Settings under the Performance Alerts section.

Email Alerts

Alert emails include:

• Metric summary — what triggered the alert, current value vs threshold
• Detail section — recent blocking chains or deadlock participants with query text, wait times, lock modes, database names, and client application
• XML attachment — full blocked_process_report.xml or deadlock_graph.xml for offline analysis

Alert Behavior

• Resolved notifications — when a condition clears (e.g., blocking ends), a "Cleared" notification fires
• Server silencing — right-click a server tab to acknowledge alerts, silence all alerts, or unsilence
• Always-on — the Dashboard alert engine runs independently of which tab is active, including when minimized to the system tray. The Lite edition's alert engine also runs regardless of tab visibility.
• Alert history — Dashboard keeps an in-memory alert log (accessible via MCP). Lite logs alerts to DuckDB (config_alert_log).
• Alert muting — create rules to suppress specific recurring alerts while still logging them. Rules match on server name, metric type, database, query text, wait type, or job name (AND logic across fields). Access via Settings → Manage Mute Rules, or right-click an alert in the Alert History tab. The context menu offers two muting options: Mute This Alert (pre-fills server + metric for a targeted rule) and Mute Similar Alerts (pre-fills metric only, matching across all servers). Muted alerts appear grayed out in alert history and are still recorded for auditability. Rules support optional expiration (1h, 24h, 7 days, or permanent).
• Alert details — right-click any alert in the Alert History tab and choose View Details to open a detail window. The window shows core alert fields (time, server, metric, value, threshold, notification type, status) plus context-sensitive details that vary by metric: query text and session info for long-running queries, job name and duration stats for anomalous agent jobs, per-wait-type breakdowns for poison waits, space usage by category for TempDB, and blocking/deadlock session counts.

---

Agent Job Monitoring

Both editions monitor currently running SQL Agent jobs and flag jobs that are running longer than expected.

Metric	How It Works
Current duration	Elapsed time since the job started
Average duration	Historical mean from successful completions in msdb.dbo.sysjobhistory
p95 duration	95th percentile from historical completions
Running long flag	Set when current duration exceeds the p95 threshold

The Full Edition collects this data via the collect.running_jobs_collector stored procedure (every 5 minutes). The Lite Edition queries msdb directly on each collection cycle. Both editions expose this data through the MCP get_running_jobs tool.

Gracefully skipped on Azure SQL Database, AWS RDS for SQL Server, and environments without SQL Server Agent.

---

MCP Server (LLM Integration)

Both editions include an embedded Model Context Protocol server that exposes monitoring data to LLM clients like Claude Code and Cursor.

Setup

1. Enable the MCP server in Settings (checkbox + port, default 5151)
   • The port must be between 1024 and 65535. Ports 0–1023 are well-known privileged ports reserved by the operating system.
   • On save, the app checks whether the chosen port is already in use and warns you if there is a conflict.
   • On startup, the app verifies the port is available before starting the MCP server.
2. Register with Claude Code:

claude mcp add --transport http --scope user sql-monitor http://localhost:5151/

3. Open a new Claude Code session and ask questions like:
   • "What servers are being monitored?"
   • "What are the top wait types on my server?"
   • "Are there any blocking or deadlock issues?"
   • "Show me CPU utilization for the last 4 hours"
   • "What are the most expensive queries by CPU?"

Available Tools

Full Edition exposes 63 tools, Lite Edition exposes 52. Core tools are shared across both editions.

Category	Tools
Discovery	list_servers
Health	get_server_summary*, get_daily_summary**, get_collection_health
Alerts	get_alert_history, get_alert_settings, get_mute_rules
Waits	get_wait_stats, get_wait_types*, get_wait_trend, get_waiting_tasks*
Queries	get_top_queries_by_cpu, get_top_procedures_by_cpu, get_query_store_top, get_expensive_queries**, get_query_duration_trend*, get_query_trend
Active Queries	get_active_queries
CPU	get_cpu_utilization
Memory	get_memory_stats, get_memory_trend, get_memory_clerks, get_memory_grants
Blocking	get_blocking**, get_deadlocks, get_deadlock_detail, get_blocked_process_reports*, get_blocked_process_xml, get_blocking_deadlock_stats**, get_blocking_trend*, get_deadlock_trend*
I/O	get_file_io_stats, get_file_io_trend
TempDB	get_tempdb_trend
Perfmon	get_perfmon_stats, get_perfmon_trend
Jobs	get_running_jobs
Configuration	get_server_config*, get_database_config*, get_database_scoped_config*, get_trace_flags*
Config History	get_server_config_changes**, get_database_config_changes**, get_trace_flag_changes**
Server Info	get_server_properties, get_database_sizes
Sessions	get_session_stats
Scheduler	get_cpu_scheduler_pressure**
Latch/Spinlock	get_latch_stats**, get_spinlock_stats**
Diagnostics	get_plan_cache_bloat**, get_critical_issues**
System Events	get_default_trace_events**, get_trace_analysis**, get_memory_pressure_events
Health Parser	get_health_parser_system_health**, get_health_parser_severe_errors**, get_health_parser_io_issues**, get_health_parser_scheduler_issues**, get_health_parser_memory_conditions**, get_health_parser_cpu_tasks**, get_health_parser_memory_broker**, get_health_parser_memory_node_oom**
Plan Analysis	analyze_query_plan, analyze_procedure_plan, analyze_query_store_plan, analyze_plan_xml, get_plan_xml
Diagnostic Analysis	analyze_server*, get_analysis_facts*, compare_analysis*, audit_config*, get_analysis_findings*, mute_analysis_finding*

* Lite only | ** Full only

Most tools accept optional server_name and hours_back parameters. If only one server is configured, server_name is auto-resolved.

The MCP server binds to localhost only and does not accept remote connections.

---

Performance Impact

On Monitored Servers

• All queries use READ UNCOMMITTED isolation
• Configurable collection intervals
• Full Edition: typical overhead <1% CPU, <100 MB memory
• Lite Edition: max 7 concurrent SQL connections, 30-second command timeout

Local Resources (Lite)

• DuckDB: ~50–200 MB per server per week
• Parquet archives: ~10x compression with ZSTD
• ScottPlot charts use hardware-accelerated rendering

---

Troubleshooting

Full Edition

Two diagnostic scripts in the install/ folder:

Script	Purpose
99_installer_troubleshooting.sql	Quick health checks: collection log errors, schedule status, Agent job status, table row counts
99_user_troubleshooting.sql	Comprehensive diagnostics: runs collectors with @debug = 1, detailed timing and row counts

SELECT
    collection_time,
    collector_name,
    error_message
FROM PerformanceMonitor.config.collection_log
WHERE collection_status = 'ERROR'
ORDER BY collection_time DESC;

Lite Edition

Application logs are written to the logs/ folder. Collection success/failure is also logged to the collection_log table in DuckDB.

Common issues:

1. No data after connecting — Wait for the first collection cycle (1–5 minutes). Check logs for connection errors.
2. Query Store tab empty — Query Store must be enabled on the target database (ALTER DATABASE [YourDB] SET QUERY_STORE = ON).
3. Blocked process reports empty — Both editions attempt to auto-configure the blocked process threshold to 5 seconds via sp_configure. On AWS RDS, sp_configure is not available — you must set blocked process threshold (s) through an RDS Parameter Group (see "AWS RDS Parameter Group Configuration" above). On Azure SQL Database, the threshold is fixed at 20 seconds and cannot be changed. If you still see no data on other platforms, verify the login has ALTER SETTINGS permission.
4. Connection failures — Verify network connectivity, firewall rules, and that the login has the required permissions. For Azure SQL Database, use a contained database user with VIEW DATABASE STATE.
5. FinOps Index Analysis hangs, times out, or returns Msg 229 on sql_expression_dependencies — sp_IndexCleanup runs against each user database under your dashboard/Lite login. If that login has no user mapping in a target database, the procedure can hang at 100% CPU instead of erroring; if it is mapped but missing SELECT on sys.sql_expression_dependencies, it errors immediately on databases that have UDF-bound computed columns or check constraints. See FinOps Index Analysis below for the full per-database grant set that fixes both.

---

Permissions

Full Edition (On-Premises)

The installer needs sysadmin to create the database, Agent jobs, and configure sp_configure settings. After installation, the collection jobs can run under a least-privilege login with these grants:

USE [master];
CREATE LOGIN [SQLServerPerfMon] WITH PASSWORD = N'YourStrongPassword';
GRANT VIEW SERVER STATE TO [SQLServerPerfMon];

USE [PerformanceMonitor];
CREATE USER [SQLServerPerfMon] FOR LOGIN [SQLServerPerfMon];
ALTER ROLE [db_owner] ADD MEMBER [SQLServerPerfMon];

USE [msdb];
CREATE USER [SQLServerPerfMon] FOR LOGIN [SQLServerPerfMon];
ALTER ROLE [SQLAgentReaderRole] ADD MEMBER [SQLServerPerfMon];

Grant	Why
VIEW SERVER STATE	All DMV access (wait stats, query stats, memory, CPU, file I/O, etc.)
db_owner on PerformanceMonitor	Collectors insert data, create/alter tables, execute procedures. Scoped to just this database — not sysadmin.
SQLAgentReaderRole on msdb	Read sysjobs, sysjobactivity, sysjobhistory for the running jobs collector

Optional (gracefully skipped if missing):

• ALTER SETTINGS — installer sets blocked process threshold via sp_configure. Skipped with a warning if unavailable.
• ALTER TRACE — default trace collector. Skipped if denied.
• DBCC TRACESTATUS — server config collector skips trace flag detection if denied.

Change the SQL Agent job owner to the new login after installation if you want to run under least privilege end-to-end.

Lite Edition (On-Premises)

Nothing is installed on the target server. The login only needs:

USE [master];
GRANT VIEW SERVER STATE TO [YourLogin];

-- Optional: for SQL Agent job monitoring
USE [msdb];
CREATE USER [YourLogin] FOR LOGIN [YourLogin];
ALTER ROLE [SQLAgentReaderRole] ADD MEMBER [YourLogin];

FinOps Index Analysis (per-database grants)

Applies to both editions. The FinOps Index Analysis tab runs sp_IndexCleanup against each user database you ask it to inspect, executing as your dashboard/Lite login. The grants above (VIEW SERVER STATE, db_owner on PerformanceMonitor, SQLAgentReaderRole on msdb) are not sufficient on their own — the login also needs a user mapping in every user database it will analyze, plus VIEW DATABASE STATE, VIEW DEFINITION, and SELECT on sys.sql_expression_dependencies in each.

The third grant is the easy one to miss: by default only members of db_owner have SELECT on sys.sql_expression_dependencies, and VIEW DEFINITION does not include it. sp_IndexCleanup queries that catalog view (via three-part name to the target database) when checking for computed columns and check constraints that reference UDFs, so the failure only surfaces on databases that actually have those — which is why a smoke-test database may pass and a real workload database fails with Msg 229.

For each target user database:

USE [YourTargetDatabase];
CREATE USER [SQLServerPerfMon] FOR LOGIN [SQLServerPerfMon];
GRANT VIEW DATABASE STATE                       TO [SQLServerPerfMon];
GRANT VIEW DEFINITION                           TO [SQLServerPerfMon];
GRANT SELECT ON sys.sql_expression_dependencies TO [SQLServerPerfMon];

Or apply broadly with sp_MSforeachdb:

EXEC sp_MSforeachdb N'
USE [?];
IF DB_ID() > 4 AND DATABASEPROPERTYEX(DB_NAME(), ''Updateability'') = ''READ_WRITE''
BEGIN
    IF NOT EXISTS (SELECT 1 FROM sys.database_principals WHERE name = ''SQLServerPerfMon'')
        CREATE USER [SQLServerPerfMon] FOR LOGIN [SQLServerPerfMon];
    GRANT VIEW DATABASE STATE                       TO [SQLServerPerfMon];
    GRANT VIEW DEFINITION                           TO [SQLServerPerfMon];
    GRANT SELECT ON sys.sql_expression_dependencies TO [SQLServerPerfMon];
END';

Symptoms if missing. There are two distinct failure modes depending on which grant is absent:

• No user mapping in the target database — sp_IndexCleanup can hang at 100% CPU with no waits and never return, instead of failing fast with Msg 916 like every other catalog DMV. The hang isn't a deadlock or a long-running scan; it's a SQL Server engine bug where a permission check at execute time gets misclassified as "this plan needs to be recompiled," producing an infinite recompile loop. Reproduces on SQL Server 2016 SP3 through 2025 CU4.
• User is mapped with VIEW DATABASE STATE + VIEW DEFINITION but no SELECT on sys.sql_expression_dependencies — fails fast with Msg 229, Level 14, State 5: The SELECT permission was denied on the object 'sql_expression_dependencies', database 'mssqlsystemresource', schema 'sys' the moment a database with a UDF-bound computed column or check constraint is reached.

Adding all three grants above eliminates both. See issue #915 for the full diagnosis.

Azure SQL Database (Lite Only)

Azure SQL Database doesn't support server-level logins. Create a contained database user directly on the target database:

-- Connect to your target database (not master)
CREATE USER [SQLServerPerfMon] WITH PASSWORD = 'YourStrongPassword';
GRANT VIEW DATABASE STATE TO [SQLServerPerfMon];

When connecting in Lite, specify the database name in the connection. SQL Agent and msdb are not available on Azure SQL Database — those collectors are skipped automatically.

Azure SQL Managed Instance

Works like on-premises. Use server-level logins with VIEW SERVER STATE. SQL Agent is available.

AWS RDS for SQL Server

Use the RDS master user for installation. The master user has the necessary permissions. For ongoing collection, VIEW SERVER STATE and msdb access work the same as on-premises, but sp_configure is not available (use RDS Parameter Groups instead — see above).

---

Folder Structure

Monitor/
│
│   Full Edition (server-installed collectors + separate dashboard)
├── install/          # 58 SQL installation scripts
├── upgrades/         # Version-specific upgrade scripts
├── Installer/        # CLI installer for Full Edition database (C#)
├── Installer.Core/   # Shared installation library (CLI + Dashboard)
├── Dashboard/        # Full Edition dashboard application (WPF, includes installer)
│
│   Lite Edition (standalone desktop app, nothing installed on server)
├── Lite/             # Lite Edition desktop application (WPF)
│
│   Shared
└── README.md         # This file

---

Building from Source

All projects target .NET 8.0.

# Full Edition Dashboard
dotnet build Dashboard/Dashboard.csproj

# Lite Edition
dotnet build Lite/PerformanceMonitorLite.csproj

# CLI Installer (self-contained)
dotnet publish Installer/PerformanceMonitorInstaller.csproj -c Release

---

Support & Sponsorship

This project is free and open source under the MIT License. The software is fully functional with no features withheld — every user gets the same tool, same collectors, same MCP integration.

However, some organizations have procurement or compliance policies that require a formal vendor relationship, a support agreement, or an invoice on file before software can be deployed to production. If that sounds familiar, two commercial support tiers are available:

Tier	Annual Cost	What You Get
Supported	$500/year	Email support (2-business-day response), compatibility guarantees for new SQL Server versions, vendor agreement and invoices for compliance, unlimited instances
Priority	$2,500/year	Next-business-day email response, quarterly live Q&A sessions, early access to new features, roadmap input, unlimited instances

Both tiers cover unlimited SQL Server instances. The software itself is identical — commercial support is about the relationship, not a feature gate.

Read more about the free tool and commercial options | Purchase a support subscription

If you find the project valuable, you can also support continued development:

Sponsor on GitHub	Become a sponsor to fund new features, ongoing maintenance, and SQL Server version support.
Consulting Services	Hire me for hands-on consulting if you need help analyzing the data this tool collects? Want expert assistance fixing the issues it uncovers?

Neither sponsorship nor consulting is required — use the tool freely.

---

Third-Party Components

sp_WhoIsActive

• Author: Adam Machanic | License: GPLv3
• Repository: https://github.com/amachanic/sp_whoisactive

DarlingData

• Author: Erik Darling (Darling Data, LLC) | License: MIT
• Repository: https://github.com/erikdarlingdata/DarlingData

SQL Server First Responder Kit

• Author: Brent Ozar Unlimited | License: MIT
• Repository: https://github.com/BrentOzarULTD/SQL-Server-First-Responder-Kit

See THIRD_PARTY_NOTICES.md for complete license texts.

---

Sponsors

Free code signing on Windows provided by SignPath.io, certificate by SignPath Foundation

---

License

Copyright (c) 2026 Darling Data, LLC. Licensed under the MIT License. See LICENSE for details.

Author

Erik Darling — erikdarling.com — Darling Data, LLC