GBase AIOps
AI-Native Intelligent Operations Platform
Core Architecture
- Health Model: Constructs a percentage-based health scoring system to evaluate the database's operating status from multiple dimensions;
- Fault Model: 800+ typical fault scenarios, covering all scenarios such as connection exhaustion, deadlock, and IO bottleneck;
- Performance Model: Supports slow SQL analysis, execution plan optimization, and resource load assessment;
- Load Model: Analyzes business traffic and concurrency pressure to support capacity planning and scaling decisions;
- Anomaly Detection Model: Based on a dynamic baseline algorithm, identifies fluctuations in indicators that deviate from normal patterns and provides early warnings of risks.
- Context sorting framework: understands operation and maintenance intent, associates historical operations, restores business context, and eliminates "conversation gaps";
- Semantic operation and maintenance knowledge graph: accumulates 200,000+ operation and maintenance experiences and 2.5 million+ knowledge questions and answers, and realizes semantic association of indicators → faults → root causes → solutions;
- SKILLS Intelligent Agent Collaboration: Parallel scheduling of multiple intelligent agents, dividing tasks such as alarms, inspections, and optimizations, realizing the decomposition and execution of complex operation and maintenance scenarios;
- Intelligent Agent Tool Pool: Built-in 2000+ operation and maintenance tools, supporting automatic invocation of tools such as SQL analysis, log investigation, and space management to automate operation and maintenance operations.
- Intelligent monitoring and early warning: real-time monitoring of database operation status, supports dynamic baseline alarms and three-level alarm hierarchical push;
- Automatic inspection: automatically performs full-dimensional inspections according to a custom cycle, and generates visual reports and optimization suggestions;
- Intelligent diagnostic analysis: based on knowledge graphs and fault models, automatically locates root causes and outputs executable handling solutions;
- Intelligent SQL Review and Optimization: Supports multi-mode SQL review, execution plan analysis, and index optimization suggestions to improve SQL performance;
- AI Natural Language Interactive Operations and Maintenance: Supports spoken language command input, enabling querying, inspection, and analysis operations through conversational interaction.
Core Capabilities
AI Natural Language Interactive Operations and Maintenance
Breaks down the barriers between traditional command-line and interface operations, supporting conversational operation and maintenance command input (such as "query and optimize GBase tablespace utilization" and "analyze the root cause of slow SQL"). Execution flow: accurate semantic recognition → contextual analysis → task decomposition → automatic tool invocation → real-time data collection → model inference analysis → result visualization output. Intent recognition accuracy is 100%, single-round interaction response time is ≤10 seconds, allowing even non-professional DBAs to complete expert-level operation and maintenance operations.
Unified management of the entire stack database
Agentless and non-intrusive, adaptable to 20+ databases including GBase 8c, GBase 8a, Oracle, and MySQL, supporting one-click management of single-machine, master-slave, and distributed clusters; unified monitoring entry point, centrally displaying running status, performance indicators, and health scores, achieving "one-screen management and full-domain visibility" for heterogeneous databases.
Intelligent Alarms and Root Cause Diagnosis
Built-in 800+ fault models, covering all scenarios of faults such as connection exhaustion, lock wait, IO bottleneck, tablespace full, and slow SQL storms; automatic execution after alarm triggering: alarm capture → indicator / session / log collection → knowledge graph reasoning → root cause precise location → handling plan generation, diagnostic accuracy >95%, no AI illusion, response ≤10 seconds, completely eliminating alarm storms and blind manual investigation.
Intelligent Optimization of the Entire SQL Lifecycle
Supports three modes: quick review, standard review, and code review; built-in 140+ review rules; automatically captures top slow SQL queries; execution flow: SQL parsing → execution plan analysis → problem identification (full table scan / missing index) → optimization suggestion generation → performance improvement estimation; one-click output of index creation and SQL rewriting solutions; optimization suggestions are implementable; single SQL optimization ≤10 seconds.
Automated Inspection and Health Management
Built-in ≥15 core inspection indicators; supports custom inspection templates and cycles; automatically executes multi-dimensional indicator collection → health score calculation (out of 100) → risk item identification → optimization suggestion generation → visual report output; automatically generates daily inspection reports, covering all dimensions of GBase database operation, performance, security, and compliance; provides early warning of potential risks; improves operation and maintenance efficiency by 50%+.
Intelligent Capacity Prediction and Risk Warning
Based on time-series algorithms and large-scale model inference, it automatically collects historical data on CPU, memory, disk, and tablespace, and executes: data modeling → trend analysis → 30-day capacity prediction → remaining available days calculation → expansion suggestion output. The prediction results closely match actual business growth, proactively avoiding storage overflow risks and ensuring continuous business operation.
O&M Knowledge Accumulation and Reuse
Build a dedicated database O&M knowledge base, automatically accumulating fault cases, optimization experience, parameter configurations, and handling processes, forming a closed loop of "experience → knowledge → capability"; supports natural language search, allowing new employees to quickly get started, and expert experience can be replicated and passed on, solving the pain point of O&M talent shortage.
Multi-mode SQL review and code-level compliance control
Supports three modes: fast review, standard review, and code review. It can perform syntax verification, performance risk assessment, and security compliance checks on SQL for deployment in the GBase database. Among them, the code review mode can directly extract SQL in batches from application code files, identify potential performance hazards and violations in advance, ensure the compatibility and stability of business code and GBase database, and avoid business failures after deployment.
Product Advantages
Deep Adaptation to the GBase Ecosystem
Fully compatible with the entire GBase 8c and GBase 8a database series, adaptable to various deployment models including distributed, master-slave, and single-machine configurations, aligning with GBase user operation and maintenance habits, and seamlessly integrated into the GBase ecosystem.
Five-layer architecture support, natively driven by AI capabilities
From the data layer to the service layer, the entire chain is designed as an AI operation and maintenance service. Through core components such as knowledge graphs, intelligent agent collaboration, and fault models, it achieves full-process automation and intelligence in operation and maintenance, rather than simply adding traditional monitoring tools.
End-to-End Security and Compliance
Localized deployment, data does not leave the domain, no external network dependence; full operation log retention, auditable, and traceable, meeting the requirements of Level 3 Information Security Protection, Information Technology Innovation Compliance, and industry security.
Lightweight and easy to deploy
Agentless and non-intrusive data collection, without modifying the database or affecting business performance (impact less than 1%); supports deployment on virtualized and China produced servers, with installation and management completed in minutes, enabling rapid deployment.
Highly scalable and easy to iterate
Supports elastic expansion of collection nodes, smoothly managing 1000+ database instances; knowledge base, fault model, and intelligent agent capabilities are continuously iterated and updated to adapt to new scenarios and new databases, protecting user investment.
Lightweight and Highly Adaptable
The platform adopts a lightweight microservice architecture, which is extremely simple to deploy, has low resource consumption, and can seamlessly connect to the GBase database and existing monitoring, ticketing, and authentication systems. It supports China-native software and hardware environments, is compatible with mainstream China developed servers and operating systems, and does not require large-scale modification of existing IT architecture. It is ready to use out of the box and can be deployed quickly, significantly reducing the launch cycle and implementation costs.
Application Scenarios
Large-Scale Operation and Maintenance of GBase Databases
Adapt to GBase distributed clusters and unified operation and maintenance of multiple branch nodes, enabling batch management, centralized monitoring, intelligent diagnosis, and automatic inspection, reducing the manpower cost of large-scale GBase database operation and maintenance.
Safeguarding China-native Database Migration
Support the entire migration process from Oracle/MySQL to GBase and China-native databases, providing pre-migration SQL review, in-migration performance monitoring, and post-migration intelligent tuning to ensure a smooth migration transition and stable performance.
Ensuring High Availability for Critical Businesses
For core businesses such as finance, government affairs, energy, healthcare, transportation, operators, and manufacturing, providing 24/7 intelligent monitoring, second-level fault diagnosis, early risk warning, and automatic handling to ensure the continuous operation of core businesses supported by GBase databases.
Compliant Operation and Maintenance in China-native Environments
Adaptable to China produced servers and Kylin V10 / UOS operating system, meeting the requirements of China-native IT innovation compliance, data security, and operation auditing, assisting government and enterprise units in China-native substitution and compliance implementation.
Business Value
Cost Reduction & Efficiency
1.Reduces operation and maintenance manpower by 50%+, shortens fault location from 2+ hours to less than 20 minutes, significantly improving operation and maintenance efficiency;
Talent Empowerment
2.Accumulates expert experience, lowers the operation and maintenance threshold, and solves the pain point of shortage of operation and maintenance talent for China-native databases;
Security & Compliance
Localized deployment, data not leaving the domain, full-link auditing, meeting the security and compliance requirements of China-native database innovation and the industry;
Ecosystem Collaboration
Deeply adapts to the entire GBase database series, improves the GBase ecosystem, and helps GBase achieve large-scale promotion;
Continuous Evolution
AI capabilities and knowledge bases are continuously iterated to adapt to new scenarios and new databases, supporting the long-term development of user businesses.
Improves the GBase ecosystem and enhances customer stickiness: As the officially recommended intelligent operation and maintenance platform of GBase, it can be directly embedded into the GBase product matrix to provide customers with end-to-end solutions from database deployment to operation and maintenance.