This topic describes how Database Autonomy Service (DAS) helps JIESHUN enable automatic database O&M.

Customer profile

JIESHUN is a provider that develops, produces, and markets comprehensive services such as smart parking and smart community services. The product line of JIESHUN includes a wide range of software and hardware products. These products include apps that provide smart parking and smart community services, hardware for smart parking management, smart access control, pedestrian access gates, a management system for city-level all-in-one smart cards, and smart platforms for parking management and online to offline (O2O) business operations.

Challenges

  • JIESHUN has difficulties in business transformation. In the early stage of business development, the parking management system and access control system developed by JIESHUN are independently deployed in an offline manner. High performance of these systems is not required. To reduce the O&M costs for customers, the parking management and access control systems support software as a service (SaaS) deployment in recent years. The SaaS services help large-sized tenants handle the business in China. This causes unexpected highly concurrent requests. Therefore, higher database performance and stability are required.
  • The insufficient capability of the original monitoring system has the following negative impacts:
    • If all databases of multiple tenants run on the same database instance, the instance resources may be insufficient to support the databases. As a result, the services related to the databases are affected.
    • If spikes in the workloads of a tenant occur, database O&M engineers cannot detect the unexpected increase in the resource utilization of the shared databases at the earliest opportunity. The data in the databases of large-sized tenants cannot be migrated to a dedicated database instance. As a result, the services of all tenants on the database instance are affected.
  • The business code migrated from the original system to the SaaS platform runs as expected in low-concurrency scenarios. However, the increase in the number of concurrent requests prolongs the response time. In particular, the order system developed by JIESHUN has poor performance in high-concurrency scenarios. This causes a large number of pending requests. As a result, traffic control in a parking lot is affected.

Solutions

  • JIESHUN connects ApsaraDB RDS database instances to DAS. This allows JIESHUN to use monitoring dashboards and the inspection and scoring feature. In addition, JIESHUN uses the anomaly detection feature to monitor the performance of database instances and the trend of disk usage. This helps JIESHUN detect potential risks. For more information, see Monitoring dashboards, Inspection and scoring, and Anomaly detection.
  • JIESHUN enables the automatic SQL throttling feature and the automatic SQL optimization feature of DAS. This way, DAS identifies slow SQL statements and throttles the concurrent requests that contain the slow SQL statements over a period of time. Then, DAS optimizes the slow SQL statements or provides optimization suggestions based on business requirements. This avoids system failures caused by resource exhaustion. When the slow SQL statements that are carried over from the original system are executed in high-concurrency scenarios, resource exhaustion occurs. For more information, see Automatic SQL throttling and Automatic SQL optimization. Automatic SQL optimization

Benefits

  • DAS allows the database administrators (DBAs) of JIESHUN to view the trend of database instances so that these DBAs can prepare for potential performance downgrades at the earliest opportunity. The system failure rate is reduced by more than 60%. This improves user experience and satisfaction.
  • DAS enables automatic SQL throttling and optimization. This allows the DBAs of JIESHUN to take more active roles as database architects. For example, the DBAs can develop and optimize the architecture of database systems, check database changes, and perform other related operations. These operations help further reduce system failures.
  • DAS provides autonomy services to manage database instances. This reduces the O&M costs of JIESHUN by more than 30%.

Customer feedback

"DAS helps JIESHUN manage dozens of database instances. The anomaly detection, automatic SQL throttling, automatic SQL optimization, and auto scaling features are used to reduce O&M costs and downtime and improve the availability of systems." -- Chief engineer and O&M engineer of JIESHUN