This topic describes the features of Database Autonomy Service (DAS).
- SQL review and optimization
- Analysis of slow SQL queries
- Storage analysis
- Performance trends
- Session management
- Full SQL statistics
- Diagnosis reports
- 24/7 anomaly detection
- Automatic SQL throttling
- Automatic SQL optimization
- Automatic SQL review and optimization
- Capacity assessment and specification recommendations
- Auto scaling
Monitoring and alerting
DAS allows you to centrally manage and monitor your on-premises databases and cloud databases. This reduces your management costs by more than 50%. The operational faults can also be significantly fewer.
- DAS provides the centralized monitoring feature. You can monitor the performance trends
and real-time performance of all the database instances in different clusters and
- Cost-effectiveness: DAS helps you monitor databases and spares you from hiring engineers to develop and deploy programs for data collection, computing, and storage.
- Various metrics: DAS collects, computes, and then displays a variety of key monitoring metrics that are related to database performance.
- Fine-grained monitoring: You can configure fine-grained monitoring to quickly detect anomalies based on your business needs. The finest granularity is seconds.
- DAS provides the centralized alerting feature. You can create custom alert rules for
on-premises databases and cloud databases and configure alert notification methods.
- Default alert templates: You can use a variety of default alert templates for different types of database engines. These default templates are developed based on the experience in database O&M that is accumulated by Alibaba Group.
- Custom configurations: DAS supports custom alert templates. You can configure alert rules, contacts, and contact groups based on your business needs. This way, you can create alert templates for different users in your enterprise.
- Smart reminders: DAS automatically checks whether alerts are configured for all the database instances that are added to it. If no alerts are configured for a database instance, DAS sends you a reminder. This protects your business against damage that can be avoided by alerting in advance.
DAS provides 24/7 anomaly detection based on machine learning and fine-grained data monitoring. You can enable a variety of features, such as automatic SQL throttling, anomaly snapshots, automatic SQL review and optimization, automatic storage expansion, and auto scaling of computing resources. These features support a closed-loop diagnosis process. This helps you detect anomalies, analyze root causes, mitigate loss or optimize performance, track effects, and roll back operations or add knowledge to the knowledge base. The effect of performance optimization can also be quantified. This way, DAS ensures the service continuity of databases.
DAS supports autonomous scenarios without manual intervention. In addition, the DAS system can continuously develop self-learning capabilities, such as automatic annotation of anomalies, case system setups, anomaly simulation, and quantitative feedback and assessment. DAS accumulates large amounts of cases based on various online business scenarios. These cases continuously accelerate DAS development so that autonomy services can be provided much more effectively. DAS has six core autonomous features based on the preceding concepts: 24/7 real-time anomaly detection, fault self-repair, automatic optimization, automatic parameter tuning, auto scaling, and intelligent stress testing.
- Data-driving capability: DAS collects large amounts of real-time data, such as performance metrics, query logs of loaded SQL statements, and logs of O&M changes. The collected data helps you develop detection capabilities to support the real-time perception of environments and threats and the real-time anomaly detection.
- Automatic decision-making capability: The deep integration of machine learning and database expert experience allows DAS to automatically make decisions in different business scenarios.
- Automatic execution capability: Tasks are automatically orchestrated and implemented based on decisions that are made by the Autonomy Center.
- Closed-loop management: A closed-loop process is provided to coordinate operations and manage services in DAS. After an anomaly is detected, DAS performs overall decision-making based on root cause analysis, followed by self-repair and performance optimization. The optimization effects are tracked and assessed. Then, DAS provides feedback or rolls back operations.
Enterprise-level database management
DAS supports various monitoring scenarios based on years of database O&M and management experience of the Alibaba Cloud database team. The database instances that DAS monitors may be deployed in the same or different clusters or environments. DAS provides multi-module dashboards for you to view performance information of all database instances.
- Management of databases in different environments or clusters
DAS allows you to monitor database instances that are deployed in multiple clusters or environments in your enterprise. You can perform aggregation and drill-down on monitoring metrics for a specific cluster or environment. Therefore, DAS meets the requirement for managing databases from an enterprise perspective.
- Inspection and scoring
DAS inspects all your database instances that are added to it. The inspection items include basic items, SQL statements, capacity, performance, and security. Based on inspection results, DAS generates a score for each database instance. The score indicates the health status of the database instance.
DAS provides services such as high-risk SQL identification, SQL injection detection, new access source identification, and discovery of access to sensitive data. All data in your databases is detected and audited in real time. DAS detects anomalies such as abnormal access to databases and data breaches to ensure the database security in a fast and effective manner.
DAS provides the following features for security audit:
- SQL injection detection
- High-risk SQL identification
- New access source identification