🦝 RaccoonX

Open Source Intelligent Database Inspection & Operations Platform

RaccoonX is an open-source, cross-platform database inspection and operations platform designed for DBAs, database engineers, DevOps teams, and infrastructure teams.

Formerly known as DBCheck.

🗄️ 22 database types 🔍 839 inspection rules 🤖 AI-assisted diagnostics 🔌 Extensible plugin architecture 📜 Apache License 2.0

If RaccoonX is useful to you, consider giving the repository a ⭐.

A Star helps more database engineers discover the project and is one of the simplest ways to support an independent open-source project.

🌐 Project Links

Website: https://raccoonx.cn/
GitHub: https://github.com/fiyo/DBCheck
Docker Hub: https://hub.docker.com/r/jackge12345/dbcheck
English: README.md
中文: README_zh.md
Issues: https://github.com/fiyo/DBCheck/issues
Discussions: https://github.com/fiyo/DBCheck/discussions

🦝 Why RaccoonX?

Database environments are becoming increasingly complex.

A typical production environment may contain:

MySQL
PostgreSQL
Oracle
SQL Server
DM8
TiDB
OceanBase
Redis
MongoDB
ClickHouse
and many other database systems.

Traditional database inspection often depends on manually executing SQL scripts, checking operating-system resources, collecting screenshots, analyzing slow SQL, and writing inspection reports.

RaccoonX tries to bring these tasks together into one open-source platform.

RaccoonX helps you:

Connect
   ↓
Inspect
   ↓
Collect
   ↓
Analyze
   ↓
Identify Risks
   ↓
Generate Reports
   ↓
Track Changes
   ↓
AI-assisted Diagnosis

It can be used as a standalone database inspection tool or as part of a broader database operations workflow.

📊 Project Milestones

RaccoonX started as DBCheck and has evolved through continuous development and community feedback.

Milestone	Status
GitHub Stars	180+
GitHub Forks	60+
Database types	22
Inspection rules	839
Languages	9
License	Apache 2.0

Docker pull counts change frequently and are shown live by the badge at the top of this page instead of being hard-coded here. They represent image pulls and should not be interpreted as the number of unique users or installations.

⭐ Help us reach the next milestone

If you have used RaccoonX, tested it, learned from it, or simply find the project interesting:

Give it a Star.

Every Single Pull
        ↓
   Keep Building
        ↓
   ⭐ Star RaccoonX

✨ Features at a Glance

Feature	Description
🗄️ Data Source Manager	Unified management of database instances, grouping, batch inspection, CSV import/export
📋 Database Inspection	22 database types and 839 inspection rules with automated Word reports
🔌 Plugin System	Independent plugin lifecycle, plugin data, templates, baselines and rules
🔍 Slow Query Analysis	Execution plans, I/O patterns, lock waits and AI-assisted analysis
🔒 Lock Diagnostics	Blocking chains, deadlocks, long transactions and remediation suggestions
📊 Index Health	Missing, redundant and long-unused index analysis
⚙️ Configuration Baseline	Compare database parameters against configurable recommended values
📈 Historical Analysis	Multi-round inspection history, trends and before/after comparisons
🤖 AI Diagnostics	Local Ollama or compatible cloud APIs for AI-assisted analysis
💬 AI Inspection Chat	Natural-language interaction with the inspection system
🧰 MCP Server	Exposes 12 tools (9 read-only + 3 approval-gated writes) to Claude, Cursor and other MCP clients
📡 Real-time Monitoring	Connections, latency, throughput and availability monitoring
🖥️ Server Inspection	CPU, memory, disk, network, services and processes
🔗 Shareable Reports	Generate report links that can be viewed without login
⏰ Scheduled Tasks	Cron-based inspection with Email/Webhook notifications
📚 RAG Knowledge Base	Upload operational documentation for AI-assisted diagnosis
📊 Oracle AWR Analysis	Parse Oracle AWR HTML reports and generate structured reports
💿 DM8 Offline Storage Check	Analyze DM8 data files without a running database
📝 SQL Editor	Interactive SQL editor with syntax highlighting and execution history
🖥️ Remote Terminal	SSH terminal with multi-tab and fullscreen support
📊 Monitoring Dashboard	Instance status matrix and topology views; connection/slow-query/capacity drill-down adapts per DB type
🧠 Intelligence Center	Multi-expert collaborative diagnosis with iterative replanning and reviewer verification
🧩 Workflow Orchestration	Visually orchestrate inspection actions (DAG) with one-click built-in marketplace templates
🛡️ Xinchuang Compliance	Compliance scoring by DB type with replacement checklist and mapping
🌐 Digital Twin	Topology snapshots, configuration drift detection, capacity and health replay
📏 Smart Baseline	Cross-instance configuration drift detection and capacity forecasting (Fleet view)
✍️ SQL Audit & Write Tickets	Write-SQL audit state machine with approval gate (WriteGate): auto-read, approved-write, full audit
💾 Disaster Recovery Backup	Scheduled database/file backups with retention and health tracking
🌍 Multi-language UI	Chinese, English, Traditional Chinese, Japanese, Korean, Spanish, French, German and Russian

🗄️ Supported Databases

RaccoonX currently supports 22 database and data systems.

Database	Driver	Default Port	Notes
MySQL	pymysql	3306	5.6 / 5.7 / 8.0+
MariaDB	pymysql	3306	10.3+
PostgreSQL	psycopg2	5432	10+
Oracle	oracledb / JPype1 + ojdbc	1521	11g R2 / 12c / 19c / 21c+, native or JDBC
SQL Server	pyodbc / JDBC	1433	2012+, native or JDBC
DM8	dmpython	5236	Dameng
TiDB	pymysql	4000	MySQL protocol
IvorySQL	psycopg2	5333	PostgreSQL compatible
YashanDB	yashandb	1688	Oracle compatible
KingbaseES	psycopg2	54321	PostgreSQL compatible
GBase 8s	JDBC	9088	JDK + JDBC
UXDB	JDBC	33060	PostgreSQL compatible
HGDB (HighGo)	JDBC	5866	PostgreSQL compatible
Vastbase G100	JDBC	5432	Vastbase, openGauss kernel
HaloDB	JDBC	5432	Xihe Data, PG/Oracle dual-compatible
GBase 8a	JDBC	5258	Column-store warehouse
MongoDB	pymongo	27017	4.0+
DB2 LUW	JDBC	50000	11.5+ / 12.x
OceanBase	pymysql	2881	MySQL tenant
TDSQL-C MySQL	pymysql	3306	MySQL compatible
Redis	redis-py	6379	3.0+, standalone and cluster
ClickHouse	JDBC	8123	21.8+

Oracle JDBC

Oracle JDBC is implemented as an independent plugin using JPype and the Oracle JDBC driver. It is useful in environments where Oracle Instant Client cannot be installed.

🚀 Quick Start

🐳 Docker — Recommended

Docker is the easiest way to start RaccoonX.

Docker Images

docker pull jackge12345/dbcheck:latest

docker run -d -p 5003:5003 \
  -v dbcheck_data:/app/data \
  -v dbcheck_reports:/app/data/reports \
  -e LD_LIBRARY_PATH=/opt/venv/lib/python3.12/site-packages/dmssl \
  jackge12345/dbcheck:latest

GitHub Container Registry

docker pull ghcr.io/fiyo/dbcheck:latest

docker run -d -p 5003:5003 \
  -v dbcheck_data:/app/data \
  -v dbcheck_reports:/app/data/reports \
  -e LD_LIBRARY_PATH=/opt/venv/lib/python3.12/site-packages/dmssl \
  ghcr.io/fiyo/dbcheck:latest

Open:

http://localhost:5003

Default credentials:

Username: admin
Password: admin123

Please change the password after the first login.

🐳 Docker Compose

curl -o deploy/docker-compose.yml \
  https://raw.githubusercontent.com/fiyo/DBCheck/main/deploy/docker-compose.yml

docker compose -f deploy/docker-compose.yml up -d

GBase 8s

The Docker image includes the required JDK and JDBC driver. GBase 8s can be used without additional driver installation when running through the provided image.

🐧 Kylin / ARM64 domestic server deployment

The official image is a multi-architecture image (linux/amd64 + linux/arm64). docker pull picks the right one automatically, so Kylin servers reporting aarch64 from uname -m (Kunpeng 920 / Phytium S2500, S5000C) work out of the box.

The image ships its own user-space runtime (Debian 12 / glibc 2.36), isolated from the host's glibc, so the host only needs kernel ≥ 4.x — Galaxy Kylin V10 SP3 ships kernel 4.19.90 and fully qualifies.

uname -m                                      # expect aarch64
# Version list: https://github.com/fiyo/DBCheck/releases
VERSION=26.11.5.1
docker pull jackge12345/dbcheck:$VERSION     # pin an explicit version
docker run -d -p 5003:5003 \
  --memory=2g --memory-swap=2g \
  -v dbcheck_data:/app/data \
  -v dbcheck_reports:/app/reports \
  --name raccoonx \
  jackge12345/dbcheck:$VERSION

Pin an explicit version tag

latest may still point at an older single-architecture image. If the pull fails with no matching manifest for linux/arm64, retry with a pinned version.

Driver differences on arm64

Database	arm64	Notes
MySQL / MariaDB / TiDB / OceanBase / TDSQL-C	✅	Pure-Python drivers, architecture independent
PostgreSQL family (IvorySQL, Kingbase, UXDB, HGDB, Vastbase, HaloDB)	✅	psycopg2 is architecture independent
DM8 (Dameng)	✅	Uses the pure-Java JDBC driver in the image, no native client needed
YashanDB	✅	Driver is a pure-Python wheel (`py3-none-any`)
MongoDB / Redis / ClickHouse	✅	Architecture independent
SQL Server	⚠️	Install msodbcsql18 yourself (the image ships only the unixodbc driver manager)
Oracle	⚠️	Thick mode needs the x86_64 Instant Client and is unavailable on arm64; thin mode and the oracle_jdbc path work
GBase 8s	⚠️	The native Python driver needs the GBase SDK — use the JDBC path on arm64

Do not run the binary package directly on Kylin

Only Windows / macOS clients are published; on Linux use the container or build from source. Debian-built binaries fail with GLIBC_2.34 not found on Kylin's glibc 2.28.

💻 Source Installation

Requirements

Python 3.10+
Git
Database-specific Python drivers
Additional JDBC / ODBC dependencies where required

Clone the repository:

git clone https://github.com/fiyo/DBCheck.git
cd DBCheck

Install dependencies:

pip install -r deploy/requirements.txt

Start the Web UI:

python web_ui.py

Open:

http://localhost:5003

CLI Mode

Chinese interface:

python -m entrypoints.cli

English interface:

python -m entrypoints.cli --lang en

Web interface:

python web_ui.py

🔍 Database Inspection

RaccoonX provides configuration-driven database inspection.

The inspection engine collects database metadata, configuration, performance information, resource usage and security-related information, then evaluates them against inspection rules.

Typical workflow:

Database
   │
   ├── Connection
   │
   ├── Metadata
   │
   ├── Configuration
   │
   ├── Performance
   │
   ├── Sessions
   │
   ├── Locks
   │
   ├── Storage
   │
   ├── Security
   │
   └── SQL
          ↓
    Inspection Engine
          ↓
     Risk Analysis
          ↓
    Recommendations
          ↓
      Word Report

📋 Inspection Coverage

RaccoonX includes database-specific inspection templates.

Typical inspection dimensions include:

Dimension	Coverage
Basic Information	Version, instance, database information
Sessions	Active sessions and connection usage
Memory	Memory configuration and utilization
Storage	Tablespaces, files and capacity
Configuration	Important database parameters
Security	Users, privileges and security configuration
SQL	Top SQL and slow queries
Locks	Blocking sessions and lock waits
Replication	Replication / Data Guard status where supported
Backup	Backup readiness and related configuration
Statistics	Object and optimizer statistics
Performance	Database performance indicators
Availability	Database and service status

Coverage varies by database type.

📊 Word Inspection Reports

RaccoonX can automatically generate structured Word inspection reports.

A typical Oracle report contains:

Chapter	Content
Cover	Database name, version, host, inspector and timestamp
Chapter 1	OS host information
Chapter 2	Database basic information
Chapter 3	Tablespaces
Chapter 4	SGA / PGA
Chapter 5	Key parameters
Chapter 6–19	Undo, Redo, Archive, DG, RAC, ASM, Sessions, Performance, Security, etc.
Chapter 20	Risks & Recommendations
Chapter 21	AI Diagnostic Suggestions
Chapter 22	Report Notes

Report chapters can be configured through the Web UI.

⚠️ Intelligent Risk Analysis

RaccoonX evaluates collected information against configurable inspection rules.

A risk item can contain:

Risk description
Current value
Recommended value
Risk level
Related SQL
Recommended action
Executable SQL where applicable

Example workflow:

Detected Risk
     ↓
Understand the Risk
     ↓
View Recommendation
     ↓
Review SQL
     ↓
Confirm
     ↓
Execute Fix

Dangerous operations such as:

DELETE
DROP
TRUNCATE

require secondary confirmation.

All execution operations are logged.

Risk recommendations are reference information. Always evaluate changes against your actual production architecture, workload and business requirements before execution.

🐢 Deep Slow Query Analysis

RaccoonX can analyze slow or expensive SQL using multiple dimensions.

Analysis may include:

SQL text
Execution plan
Execution time
CPU usage
I/O behavior
Lock waits
Index usage
Execution frequency
Resource consumption
Historical information
AI-assisted diagnosis

The goal is to connect SQL symptoms with database and system evidence rather than looking at SQL text alone.

🔒 Lock & Blocking Diagnostics

Lock analysis provides information such as:

Blocking sessions
Blocked sessions
Blocking chains
Lock waits
Long transactions
Deadlock statistics
Session relationships
Suggested remediation actions

For supported databases, remediation SQL can be generated directly from the risk analysis interface.

📊 Index Health Analysis

RaccoonX analyzes index-related conditions including:

Potentially missing indexes
Redundant indexes
Long-unused indexes
Index statistics
Index-related SQL performance issues

Recommendations should always be reviewed against actual workload characteristics before applying changes.

⚙️ Configuration Baseline

RaccoonX provides configurable baseline management.

The Web UI can be used to define:

Recommended values
Thresholds
Compliance rules
Database-specific configuration
Inspection rules

Examples include:

MySQL
PostgreSQL
Oracle
SQL Server
DM8
TiDB
YashanDB
KingbaseES
GBase 8s
MongoDB
ClickHouse

The baseline engine compares actual database parameters with configured recommendations.

📈 Historical Trend Analysis

Inspection results can be retained locally and analyzed over multiple inspection rounds.

Historical analysis can show:

Resource trends
Configuration changes
Risk changes
Performance changes
Before / after comparisons
Risk evolution over time

This makes it possible to move from:

One-time Inspection

to:

Continuous Database Health Tracking

📡 Real-time Monitoring

The Web UI provides real-time monitoring for supported database types.

Typical metrics include:

Response latency
QPS / TPS
Active connections
Total connections
Running sessions
Availability
Slow queries
Active connection heatmaps

The monitoring dashboard automatically adapts to the capabilities of each database type.

For database types without deep metrics support, RaccoonX can provide TCP-level connectivity information including:

Reachable / unreachable timeline
Availability percentage
Authentication failure
Port unreachable
Circuit-breaker state
Unsupported deep-collection status

🔥 Slow Query & Connection Heatmap

The monitoring interface provides heatmap visualization for:

Slow queries
Active connections
Time distribution
Connection activity

Auto-refresh can be configured between:

5s – 60s

CSV export is also supported.

🤖 AI Smart Diagnostics

RaccoonX supports AI-assisted database diagnostics.

The AI layer can analyze inspection results and provide:

Risk explanations
Possible root causes
Optimization suggestions
SQL analysis
Configuration suggestions
Operational recommendations

Local AI with Ollama

For environments where database information should remain local:

ollama pull qwen3:30b
ollama pull nomic-embed-text

Then start RaccoonX:

python web_ui.py

Configure the AI backend from the AI Settings page.

Supported AI Backends

Backend	Description
`ollama`	Local AI deployment
`openai`	OpenAI-compatible cloud APIs
`disabled`	Disable AI

When using Ollama, inspection data can remain within the local environment.

AI-generated recommendations should be reviewed by qualified database engineers before being applied to production systems.

💬 AI Chat Inspection

RaccoonX includes an AI interaction panel in the Web UI.

Instead of navigating through every inspection option manually, users can interact with the inspection workflow using natural language.

For example:

Show me the databases with the highest connection usage.

Analyze the slow SQL of this instance.

What are the highest-risk configuration items?

Explain why this database has high I/O.

Compare the current inspection with the previous inspection.

Available capabilities depend on the configured database and AI backend.

📚 RAG Knowledge Base

RaccoonX includes a local knowledge-base capability.

Supported document types include:

PDF
Word
Markdown
TXT

Documents can be vectorized and retrieved during AI diagnosis.

Typical workflow:

Upload Documentation
        ↓
Vectorization
        ↓
Knowledge Retrieval
        ↓
Database Inspection
        ↓
AI Diagnosis
        ↓
Context-aware Suggestions

This allows organizations to combine database inspection data with their own operational documentation and standards.

🔌 Plugin Architecture

RaccoonX uses an extensible plugin architecture.

Plugins can manage their own:

Lifecycle
Metadata
Inspection templates
Baselines
Rules
Plugin-specific data

Typical plugin structure:

plugins/available/your_plugin/
├── plugin.json
├── main_plugin.py
├── template_data.json
├── baseline_data.json
└── rules/

A plugin can be installed, enabled, disabled and uninstalled through the Web UI.

Built-in Plugins

Plugin	Database	Description
MongoDB	MongoDB 4.0+	Connection status, database statistics, slow queries
Oracle JDBC	Oracle 11g / 12c / 19c / 21c+	JDBC-based Oracle inspection
DB2 JDBC	DB2 LUW 11.5+ / 12.x	JDBC inspection and system catalog analysis
Redis	Redis 3.0+	Memory, clients, persistence, replication, security
Redis Cluster	Redis Cluster	Topology, slots, nodes and failover
UXDB JDBC	UXDB 2.x	PostgreSQL-compatible inspection
HGDB JDBC	HGDB V9	PostgreSQL-compatible inspection
TDSQL-C MySQL	TDSQL-C	MySQL-compatible inspection

Plugin development documentation:

docs/plugin/

🖥️ Server Inspection

Server inspection is independent of database inspection.

It can collect:

CPU
Memory
Disk
Network
Processes
Services
System resources

A separate server inspection report can be generated.

This allows database and operating-system information to be considered together during troubleshooting.

🖥️ Remote Terminal

RaccoonX includes an SSH-based remote terminal.

Supported features include:

Password authentication
SSH key authentication
Multiple terminal tabs
Fullscreen mode

This can be useful when database diagnosis requires checking the underlying host.

🔗 Shareable Reports

Inspection reports can be shared through generated links.

Supported capabilities include:

One-click sharing
Viewing without login
Permission isolation
Visit counting
Immediate deletion

Example:

/share/<share_id>

⏰ Scheduled Inspections

RaccoonX supports scheduled inspection tasks based on Cron expressions.

Common schedules include:

Daily
Weekdays
Weekly
Monthly
Custom Cron

After an inspection is completed, notifications can be sent through:

Email
Webhook
WeCom
DingTalk
Custom JSON Webhook

Word inspection reports can be attached to email notifications.

💾 Disaster Recovery Backup

RaccoonX includes a disaster recovery backup module.

Supported targets include:

MySQL
MariaDB
PostgreSQL
Files

Features include:

Scheduled backups
Cron scheduling
Retention cleanup
Backup history
Backup health scoring
One-click restore points
Webhook notifications
Email notifications

Database passwords are encrypted at rest using Fernet and masked in API responses.

📊 Oracle AWR Analysis

Oracle AWR HTML reports can be uploaded to RaccoonX.

The system can parse key performance information and generate a structured Word analysis report.

AI-assisted analysis can also be enabled.

Typical workflow:

Oracle AWR HTML
       ↓
Upload
       ↓
Parse
       ↓
Analyze
       ↓
Generate Report
       ↓
AI-assisted Diagnosis

💿 DM8 Offline Storage Check

RaccoonX supports DM8 storage inspection without a running database instance.

It can directly inspect:

.DBF
dm.ctl

Supported modes:

Local directory
Remote SSH server

The storage scanner can identify suspicious blocks based on binary patterns.

Examples include:

ZERO_PAGE
CONSTANT_FILL
TRUNCATED

Detected blocks are reported by:

Physical page number
File offset
Tablespace

A structured Word report can also be generated.

📝 SQL Editor

RaccoonX includes an interactive SQL editor in the Web UI.

Features include:

SQL syntax highlighting
Database object browsing
Result tables
Execution history
Friendly error messages

The editor supports the database types available through the configured RaccoonX drivers and plugins.

🌍 Multi-language Support

RaccoonX currently supports:

中文
English
繁體中文
日本語
한국어
Español
Français
Deutsch
Русский

Language can be changed from:

Web UI language selector
CLI argument

Example:

python -m entrypoints.cli --lang en

The localized areas include:

UI text
Menus
Reports
AI diagnostic labels

RaccoonX also supports:

Dark theme
Light theme

🔌 REST API

RaccoonX provides REST APIs for automation and integration.

API Key authentication can be used to integrate RaccoonX with:

CI/CD
Monitoring platforms
Automation systems
Internal operation platforms

Health Check

curl http://localhost:5003/api/v1/health

Trigger Inspection

curl -X POST http://localhost:5003/api/v1/inspect \
  -H "X-API-Key: YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "db_type": "mysql",
    "host": "192.168.1.100",
    "port": 3306,
    "user": "root",
    "password": "****"
  }'

Async Inspection

curl -X POST http://localhost:5003/api/v1/inspect \
  -H "X-API-Key: YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "db_type": "oracle",
    "host": "192.168.1.200",
    "service_name": "ORCL",
    "user": "system",
    "password": "****",
    "mode": "async"
  }'

API Endpoints

Endpoint	Method	Description
`/api/v1/health`	GET	Health check
`/api/v1/inspect`	POST	Trigger inspection
`/api/v1/inspect/{task_id}`	GET	Query task result
`/api/v1/inspects`	GET	Recent task list
`/share/<share_id>`	GET	View shared report

For production environments, use a reverse proxy such as nginx and rotate API keys regularly.

🧰 MCP Server

RaccoonX ships a Model Context Protocol (MCP) server that hands inspection capabilities directly to MCP-capable clients such as Claude, Cursor and Codex.

The point is simple: your AI no longer needs you to paste SQL output, screenshots or report contents. It can list your data sources, run an inspection, inspect blocking chains, check index health — and then reason about the result.

Start it

python -m modules.mcp_server

It uses stdio transport by default (local trusted mode). The MCP client spawns the process, so no port needs to be exposed.

Client configuration

Add this to the mcpServers block of Claude Desktop, Cursor or any MCP client:

{
  "mcpServers": {
    "dbcheck": {
      "command": "python",
      "args": ["-m", "modules.mcp_server"],
      "cwd": "/path/to/DBCheck"
    }
  }
}

Available tools (12: 9 read-only + 3 write)

Tool	Description
`dbcheck.list_instances`	List data sources visible to the current user (passwords always stripped)
`dbcheck.run_inspection`	Run one inspection round against a data source and return the verdict
`dbcheck.slow_queries`	Top slow queries, execution plans and AI optimization advice
`dbcheck.lock_tree`	Blocking chains, deadlocks and long-transaction trees
`dbcheck.index_health`	Missing, redundant and long-unused indexes
`dbcheck.baseline_check`	Compare parameters against the configuration baseline
`dbcheck.ai_diagnose`	Comprehensive AI diagnosis for one instance
`dbcheck.nl2sql`	Natural language to SQL (bridged to Chat2DB as an upstream MCP server)
`dbcheck.bicqa_ask`	Knowledge-base Q&A backed by BIC-QA
`dbcheck.kill_session`	Write. Terminate a blocking session — returns a pending-approval ticket, never executes directly
`dbcheck.apply_index`	Write. Create/rebuild an index — pending-approval ticket, revertible via `DROP INDEX`
`dbcheck.execute_sql`	Write. Run arbitrary SQL — blocked by audit rules, otherwise pending approval, then controlled execution

Write operations never execute over MCP

The three write tools do not touch the database directly. Each call is converted into a SQL statement and submitted to the same SQL audit state machine the Web UI uses (WriteGate: auto-read / approved-write / full audit). You get back APPROVAL_REQUIRED plus a task number; statements that hit a blocking rule come back as BLOCKED_BY_RULE and never enter the approval queue. An MCP client cannot name its own approver — approval and execution can only happen in the Web UI, so an AI cannot bypass approval and modify a database directly.

Tools reuse the same user-visibility checks and audit logging as the Web UI, so the permission model is identical: over MCP an ordinary user can only see the data sources they are allowed to see.

📦 Distribution Packaging

RaccoonX can be packaged as a standalone executable using PyInstaller.

Windows

rd /s /q build dist __pycache__
pyinstaller dbcheck.spec
cd dist
dbcheck.exe

Linux

pyinstaller build/dbcheck_linux.spec
cd dist
./dbcheck

🧩 Environment Quick Reference

Database	Python Driver	Extra Dependencies
MySQL / TiDB	pymysql	—
PostgreSQL / IvorySQL / KingbaseES	psycopg2-binary	—
Oracle	oracledb	No Instant Client required
SQL Server	pyodbc	ODBC Driver 17
DM8	dmpython	DM8 client libraries
YashanDB	yashandb	—
GBase 8s	jaydebeapi + JPype1	JDK + JDBC driver
Oracle JDBC	JPype1 + ojdbc	JDK + ojdbc
MongoDB	pymongo	—
DB2 LUW	JPype1 + db2jcc4	JDK + db2jcc4.jar
OceanBase	pymysql	—
Redis / Redis Cluster	redis-py	—
ClickHouse	JDBC	JDK + JDBC driver

❓ FAQ

Some inspection sections appear empty. Why?

Inspection coverage varies by database type.

RaccoonX uses database-specific templates and graceful fallback mechanisms when certain metrics are unavailable.

Connection failed. What should I check?

Check:

Database network accessibility
Firewall rules
Database listener/service status
Username and password
User privileges
Database-specific connection parameters

GBase 8s reports "Driver not found".

When using the Docker image, the required JDK and JDBC driver are already included.

For source installation, verify the JDBC driver and JDK configuration.

AI diagnostics are not working.

Check:

ollama serve

and verify that the selected model has been downloaded:

ollama pull qwen3:30b

Then verify the AI settings in the Web UI.

Oracle reports ORA-01017.

Verify:

Username
Password
Service name
Authentication mode

For SYSDBA users, enable the SYSDBA option in the Web UI or use:

sys as sysdba

Are the recommended fixes safe to execute automatically?

No recommendation should be blindly applied to production.

RaccoonX provides inspection results and suggested remediation actions. Database engineers should review the proposed change against the actual environment before execution.

🔄 From DBCheck to RaccoonX

RaccoonX was previously known as DBCheck.

The project began as a database inspection tool and gradually expanded into a broader database operations platform.

The name RaccoonX represents the next stage of the project.

DBCheck
   │
   ├── Database Inspection
   ├── Risk Analysis
   ├── Performance Analysis
   ├── AI Diagnostics
   ├── Monitoring
   ├── Plugins
   └── Automation
          ↓
      RaccoonX

The original DBCheck name may still appear in:

Repository URLs
Docker image names
Existing deployment configurations
Historical documentation
Existing scripts

These will be gradually unified as the RaccoonX brand evolves.

RaccoonX — formerly known as DBCheck.

🤝 Contributing

RaccoonX is an open-source project and contributions are welcome.

You can contribute through:

⭐ Star the repository
🐛 Report bugs
💡 Submit feature requests
🗄️ Improve database inspection rules
🔌 Develop plugins
📝 Improve documentation
🌍 Improve translations
🔧 Submit pull requests
📢 Share RaccoonX with other database engineers

Before submitting a pull request, please read:

CONTRIBUTING.md

🐛 Issues & Feature Requests

Please use GitHub Issues for:

Bug reports
Feature requests
Database compatibility problems
Installation problems
Performance issues
Documentation problems

When reporting a database-specific issue, please provide:

Database type
Database version
RaccoonX version
Operating system
Installation method
Relevant error message
Steps to reproduce

Please remove passwords, credentials, IP addresses and other sensitive information before posting.

💬 Community

Discussion topics are welcome, including:

Database operations
DBA automation
Database inspection
Database performance
AI-assisted database operations
Plugin development
Open-source development
Database compatibility

Use GitHub Discussions for questions and ideas that do not belong in a bug report.

🙏 Acknowledgements

RaccoonX references and builds upon ideas, tools and open-source projects from the database community.

Special thanks to the following projects:

We appreciate the work of everyone who contributes to the database and open-source communities.

❤️ Support the Project

RaccoonX is released under the Apache License 2.0.

If RaccoonX has been useful to you, there are many ways to support the project.

⭐ Give it a Star

A Star helps more database engineers discover RaccoonX.

Use it
  ↓
Like it
  ↓
⭐ Star it
  ↓
Share it
  ↓
More people discover it
  ↓
More feedback
  ↓
Better RaccoonX

🐛 Report a Bug

A good bug report can be more valuable than a Star.

💡 Suggest a Feature

Your real-world requirements help determine what gets built next.

🔧 Contribute Code

Pull requests are welcome.

📢 Share the Project

If you know a DBA or database engineer who might find RaccoonX useful, sharing the project is one of the most valuable forms of support.

☕ Sponsorship

Some users have supported the project through donations.

The amount of a donation is not the most important thing.

For an independent open-source project, a small donation means:

Someone used it. Someone noticed it. Someone believed it was worth continuing.

Thank you to everyone who has supported RaccoonX through donations, Stars, Issues, Pull Requests, suggestions, articles, testing, or simply telling another engineer about the project.

Please specify your name or nickname when sponsoring ❤️

📜 License

RaccoonX is open-source software licensed under the:

Apache License 2.0

See:

LICENSE

for the complete license text.

You are free to use, modify, distribute and build upon the project according to the terms of the Apache License 2.0.

⚠️ Third-Party Trademarks

The names, logos, trademarks and database technologies mentioned in this project belong to their respective owners.

Their appearance in RaccoonX indicates compatibility or integration with the corresponding technology and does not imply endorsement, affiliation or partnership.

🦝 Keep Building

RaccoonX is not built by a large team.

It grows through code, feedback, issues, ideas, testing, documentation, contributions and the people who choose to use it.

One Pull at a Time,
We Keep Building.

🦝 RaccoonX

Thank you for being part of the RaccoonX community.

Author: Jack Ge Project: RaccoonX — formerly DBCheck Website: https://raccoonx.cn/ Email: sdfiyon@gmail.com

Name		Name	Last commit message	Last commit date
.github		.github
assets		assets
build		build
data/reports		data/reports
db		db
deploy		deploy
docs		docs
drivers		drivers
experiments		experiments
i18n		i18n
modules		modules
plugins		plugins
scripts		scripts
snapshot		snapshot
templates		templates
web_templates		web_templates
.dockerignore		.dockerignore
.env.example		.env.example
.gitattributes		.gitattributes
.gitignore		.gitignore
AGENTS.md		AGENTS.md
CHANGELOG.md		CHANGELOG.md
LICENSE		LICENSE
NOTICE		NOTICE
README.md		README.md
README_zh.md		README_zh.md
diagnose_mysql_report.py		diagnose_mysql_report.py
docker-stats.json		docker-stats.json
mcp_server_config.json		mcp_server_config.json
pytest.ini		pytest.ini
web_ui.py		web_ui.py

Date	Name	ID
2026-04-28	自由的风	No.000001
2026-04-29	黄嵘	No.000002
2026-05-04	张佰政	No.000003
2026-06-02	残酷月光	No.000004
2026-06-03	大树	No.000005
2026-06-07	岳彩波（Adil0518）	No.000006
2026-06-17	轩	No.000007
2026-06-18	卿云	No.000008
2026-06-18	yuanlnet	No.000009
2026-06-18	赵法威	No.000010
2026-06-19	类延良	No.000011
2026-06-19	渺渺兮予怀	No.000012
2026-09-06	leon	No.000013
2026-09-24	James.Yao	No.000014
2026-10-10	acdante-zhang	No.000015

Latest commit

History

Folders and files

Repository files navigation

🦝 RaccoonX

Open Source Intelligent Database Inspection & Operations Platform

🌐 Project Links

🦝 Why RaccoonX?

RaccoonX helps you:

📊 Project Milestones

⭐ Help us reach the next milestone

✨ Features at a Glance

🗄️ Supported Databases

🚀 Quick Start

🐳 Docker — Recommended

Docker Images

GitHub Container Registry

🐳 Docker Compose

🐧 Kylin / ARM64 domestic server deployment

💻 Source Installation

Requirements

CLI Mode

🔍 Database Inspection

📋 Inspection Coverage

📊 Word Inspection Reports

⚠️ Intelligent Risk Analysis

🐢 Deep Slow Query Analysis

🔒 Lock & Blocking Diagnostics

📊 Index Health Analysis

⚙️ Configuration Baseline

📈 Historical Trend Analysis

📡 Real-time Monitoring

🔥 Slow Query & Connection Heatmap

🤖 AI Smart Diagnostics

Local AI with Ollama

Supported AI Backends

💬 AI Chat Inspection

📚 RAG Knowledge Base

🔌 Plugin Architecture

Built-in Plugins

🖥️ Server Inspection

🖥️ Remote Terminal

🔗 Shareable Reports

⏰ Scheduled Inspections

💾 Disaster Recovery Backup

📊 Oracle AWR Analysis

💿 DM8 Offline Storage Check

📝 SQL Editor

🌍 Multi-language Support

🔌 REST API

Health Check

Trigger Inspection

Async Inspection

API Endpoints

🧰 MCP Server

Start it

Client configuration

Available tools (12: 9 read-only + 3 write)

📦 Distribution Packaging

Windows

Linux

🧩 Environment Quick Reference

❓ FAQ

Some inspection sections appear empty. Why?

Connection failed. What should I check?

GBase 8s reports "Driver not found".

AI diagnostics are not working.

Oracle reports ORA-01017.

Are the recommended fixes safe to execute automatically?

🔄 From DBCheck to RaccoonX

🤝 Contributing

🐛 Issues & Feature Requests

💬 Community

🙏 Acknowledgements

❤️ Support the Project

⭐ Give it a Star

🐛 Report a Bug

💡 Suggest a Feature

🔧 Contribute Code

📢 Share the Project