Architecture
Google Cloud tools behind CanadianBankNews.
This diagram maps the actual platform services: public caching, Cloud Run frontend and APIs, Vertex AI, Cloud SQL, scheduled jobs, logging, analytics, and the Dev-to-Production release path.
Public frontend
Cloud CDN and Cloud Run serve public pages, guides, calculators, real-estate decision pages, Interview Prep, and credit-card tools.
AI and APIs
Cloud Run APIs coordinate advisor requests, alerts, practice interviews, analytics events, and data access. Vertex AI powers advisor answers and summaries.
Data and operations
Cloud SQL, public caches, Cloud Scheduler, Cloud Logging, Cloud Build, and Artifact Registry keep the site reproducible and observable.
Google Cloud Architecture
GCP tools, data flow, AI services, and release flow
Public requests enter through Cloud CDN and Cloud Run. Product actions call Cloud Run APIs, Vertex AI, Cloud SQL, and analytics. Releases move from GitHub to Cloud Build to Artifact Registry, then Dev verification and Production promotion use the same tested image digest.
Dev and Production use separate Cloud Run services and revisions. The intended promotion path is Git commit to Cloud Build to immutable Artifact Registry image digest, verify in Dev, then promote that same tested digest to Production without rebuilding.
Cache Safety
Cacheable by design
Visitor-independent public pages and selected public APIs can use shared cache headers: home, news, credit-card guides, real-estate guides, Interview Prep, sitemap, robots, rates, and cached news.
Never shared-cached
Admin pages, authenticated routes, saved scenarios, user-specific API responses, no-store article variants, and sensitive form submissions remain private or uncached.