All four operations are supported management actions for Huawei Cloud RDS instances, subject to the applicable database engine, instance type, current status, and other service constraints.

Changing the failover priority is supported for primary/standby RDS deployments. Huawei allows administrators to prioritize either Reliability , which emphasizes data consistency, or Availability , which emphasizes service continuity during failover.

Deleting a DB instance is a standard lifecycle-management operation available to appropriately authorized users. Huawei's RDS permission model includes the rds:instance:delete action for this purpose.

Scaling up storage space is likewise an established RDS modification operation. Administrators can increase allocated storage when database capacity requirements grow, and supported database engines can also provide storage autoscaling options.

Finally, rebooting a DB instance is a supported administrative action used when configuration changes, troubleshooting, or other operational circumstances require an instance restart. Huawei's CTS integration explicitly identifies both scaling storage and restarting an instance as auditable RDS operations.

Huawei's RDS API catalog lists deleting, scaling storage, rebooting, and changing failover priority as supported instance-management operations.

Reference topics: RDS Instance Management; Failover Priority; Storage Scaling; DB Instance Reboot; DB Instance Deletion.

1 → Bare Metal Server (BMS)

2 → Elastic Cloud Server (ECS)

3 → Physical server

The first scenario maps to BMS because Bare Metal Servers combine exclusive access to physical resources with cloud-oriented provisioning and management. Huawei Cloud recommends BMS for key or mission-critical applications and services requiring high performance. BMS retains direct physical CPU and memory access while supporting capabilities such as automatic provisioning, VPC networking, image deployment, monitoring, and cloud storage. It therefore provides both resource exclusivity and cloud flexibility .

The second scenario maps to ECS . ECS instances operate on virtualized infrastructure, so physical server resources can be shared among virtual machines. This model is appropriate for general-purpose and specialized cloud workloads where virtualization, rapid deployment, elasticity, and efficient resource utilization are priorities.

The third scenario maps to a traditional physical server . A physical server also provides exclusive resources, but Huawei's comparison shows that traditional servers do not inherently provide cloud capabilities such as automatic provisioning, VPC integration, image-based deployment, and standardized cloud monitoring. This makes them characteristic of traditional dedicated-service environments.

Huawei's comparison explicitly records exclusive resources for BMS and physical servers, but not ECS.

Reference topics: BMS, Physical Server, and ECS Comparison; Resource Exclusivity; Automatic Provisioning; Compute Architecture.

===============

Option D is incorrect . Huawei Cloud's official RDS for PostgreSQL constraints specify that a maximum of five read replicas can be created for one DB instance , not 15. Huawei's current read-replica documentation reiterates the same five-replica limit. The “15 read replicas” characteristic is associated with other Huawei database architectures, such as TaurusDB/GaussDB-related scaling designs, and must not be transferred to standard RDS for PostgreSQL.

Option C is directly supported by Huawei documentation. RDS for PostgreSQL is an object-relational database designed for complex OLTP and supports NoSQL and GIS data types. Huawei explicitly identifies websites, location-based applications, and complex data object processing as representative workloads.

Option B is also valid. Huawei Cloud provides real-time status and performance monitoring for RDS instances, exposing database-related metrics and allowing alarms to be configured for operational monitoring.

Option A reflects the managed-service security architecture: RDS provides controls involving network isolation, security groups, SSL, account authorization, backups, and other mechanisms to protect databases and user data.

The numerical scaling claim in D therefore makes it decisively incorrect.

Reference topics: RDS for PostgreSQL; Read Replicas; Database Monitoring; PostgreSQL Application Scenarios; RDS Security.

===============