What's changed: Created Professional Cloud Developer Chapter 1 (Domain 1 "Design" part 1: high-performing apps/APIs = CE/GKE/Cloud Run/Functions, LB/session affinity, Memorystore, REST/gRPC, Apigee/API Gateway, Pub/Sub/Eventarc/Cloud Tasks/Workflows, traffic splitting; secure apps = Secret Manager/Cloud KMS, service accounts/Workload Identity Federation/ADC, least privilege, IAP, Binary Authorization, Artifact Analysis/SCC, Cloud Service Mesh).
1.1Designing high-performing applications and APIs
Understand choosing platforms by use case (Compute Engine, GKE, Cloud Run), load balancers and session affinity, caching (Memorystore), creating APIs (REST, gRPC) and managing them (Apigee, API Gateway), event-driven integration (Pub/Sub, Eventarc), and traffic splitting for gradual rollouts.
A Professional Cloud Developer builds scalable, secure, reliable cloud-native applications with Google-recommended tools. First, cover choosing the right platform and building high-performing APIs.
1.1.1Choosing a platform
Choose the platform by requirements: VM control = Compute Engine, container orchestration/portability = GKE, serverless containers with minimal ops = Cloud Run, small event-driven functions = Cloud Functions. Distribute load with load balancers, use session affinity for stateful connections, and cache frequent reads with Memorystore (Redis/Memcached). Map "containers minimal ops = Cloud Run" and "speed up via cache = Memorystore."
1.1.2APIs and event-driven integration
Build APIs with widely used REST (HTTP) or low-latency, typed gRPC. Centralize API publishing, auth, rate limiting, and metering with Apigee or API Gateway. Integrate services with loosely coupled async/event-driven patterns (messaging Pub/Sub, event delivery Eventarc, tasks Cloud Tasks, orchestration Workflows). For gradual rollout of a new service, use traffic splitting (Cloud Run/GKE) for canary and rollback. Map "API management/rate limiting = Apigee/API Gateway" and "loose coupling = Pub/Sub/Eventarc."
Common: requirement → means. E.g., "serverless containers, minimal ops" = Cloud Run; "speed up frequent reads" = Memorystore; "API rate limiting/auth/metering" = Apigee/API Gateway; "loosely couple services" = Pub/Sub/Eventarc; "gradual rollout" = traffic splitting; "low-latency typed API" = gRPC.
Watch the mix-ups: (1) Choose synchronous APIs (REST/gRPC) vs async/event-driven (Pub/Sub/Eventarc) by requirements. (2) Caching (Memorystore) is for read speedup, not a durable store. (3) Traffic splitting is the means for gradual rollout/rollback.
1.1.3Section summary
- Platform = CE/GKE/Cloud Run/Functions by requirements; LB, session affinity, Memorystore
- APIs = REST/gRPC; manage via Apigee/API Gateway (auth, rate limiting, metering)
- Loose coupling = Pub/Sub/Eventarc/Cloud Tasks/Workflows; gradual rollout = traffic splitting
Sign in to track progress — Log in.
Quick check
(just a quick review)Q1. To run containers serverless with minimal ops and scale to zero when idle, which is best?
Q2. To cache frequently read data and reduce latency, which is best?
Q3. To centrally manage API auth, rate limiting, and metering, which is best?
Q4. To integrate services loosely with async/event-driven patterns, which is best?
Q5. To release a new service to a fraction of traffic first and roll back on issues, what do you use?
Q6. To build a low-latency, typed API for internal service-to-service calls, which is best?