Question

Difficulty: MediumDesigning Infrastructure for Technical Requirements and High Availability

An IoT smart-agriculture enterprise is migrating its fleet telemetry ingestion pipeline to Google Cloud. The architecture must process high-volume stateless HTTPS telemetry requests and store transactional metadata in a relational database within a single region. Technical constraints require high availability across multiple zones, zero data loss (RPO = 0), an RTO under 1 minute for database failover, and minimal operational management overhead. Which TWO architectural recommendations should you implement? (Select TWO.)

  1. Deploy the stateless HTTP telemetry ingestion workload on Cloud Run configured across multiple zones within the region.Answer
  2. B
    Deploy a regional Google Kubernetes Engine (GKE) cluster with multi-zonal node pools to host the stateless HTTP ingestion services.
  3. C
    Provision Cloud Spanner configured as a regional instance to store the transactional metadata.
  4. Provision Cloud SQL for PostgreSQL with High Availability (HA) enabled, utilizing regional persistent disks and automated zonal failover.Answer
  5. E
    Configure HA VPN with dynamic BGP routing to connect remote telemetry aggregators requiring continuous 10 Gbps throughput into the VPC.

Answer

Deploy the stateless HTTP telemetry ingestion workload on Cloud Run, and provision Cloud SQL for PostgreSQL with High Availability (HA) using regional persistent disks.
Cloud Run automatically serves containerized stateless workloads across multiple availability zones within a region while eliminating infrastructure setup and operational management. Cloud SQL configured for High Availability (HA) uses synchronous disk replication across zones in the same region, fulfilling the zero RPO requirement and providing automated failover with an RTO of less than 60 seconds.

Step-by-Step Solution

1
Evaluate stateless compute platform based on operational efficiency
Cloud Run automatically scales across multiple availability zones without cluster maintenance, satisfying stateless HTTP processing with minimal overhead.
GKE cluster management is overhead-heavy for simple stateless web endpoints.
2
Evaluate database requirements for single-region HA, RTO, and RPO
Cloud SQL with HA configuration synchronously replicates data to a standby zone using regional persistent disks, achieving RPO = 0 and automatic failover within 60 seconds.
Cloud Spanner introduces excessive cost for single-region relational workloads.

Key Concept

Designing single-region high-availability architectures for stateless compute and transactional databases on Google Cloud
Rate this question