Server-Side Container Architecture

Server-side container architecture is more powerful than client-side but requires deliberate design. This module covers the hosting, the components, and the architecture decisions that determine cost and capability.

1. Server-side container architecture

A server-side container is a server (or serverless function) running Google Tag Manager's server logic. It receives HTTP requests from clients, processes them through tags, and sends data to destinations.

2. Cloud hosting

• Google Cloud Run: The reference platform; Google's recommended setup.
• Stape: Managed Cloud Run with simpler UI.
• AWS / Azure: Possible via Docker container.
• Self-hosted: On VPS or on-prem (more setup, more control).

3. Subdomain setup

Best practice: route requests through your own subdomain (analytics.yourdomain.com or similar). Benefits:

• First-party cookie domain.
• Avoids ad-blocker domain blocklists.
• Improves data persistence in Safari.
• Brand control over the request URL.

4. Container objects

• Clients: Listen for incoming requests (GA4 Client, Web Container client, custom clients).
• Triggers: Same concept as web container, server-side.
• Variables: Server-side data references.
• Tags: Send data to destinations.

5. The GA4 and Web Container clients

The GA4 client receives data from gtag.js / GA4 events. The Web Container client receives data from a client-side tag manager. Most production setups use both.

6. Custom clients

Custom clients enable bespoke event collection. Example: a backend system sends purchase events server-to-server; a custom client receives them and forwards to vendors.

7. Data enrichment

Server-side processing can enrich events with: server-side user IDs, customer data from CDP, normalized product data, hashed PII for vendor matching.

8. Filtering and redaction

• Drop events that should not be sent to specific vendors.
• Strip PII before forwarding.
• Apply consent-mode rules server-side.
• Normalize event names across vendors.

9. Logging and monitoring

• Cloud Run logs for debugging.
• Vendor delivery success rates.
• Event volume monitoring.
• Latency monitoring.
• Alerts on data anomalies.

10. Cost optimization

• Cloud Run autoscaling: pay only when handling requests.
• Min-instance tuning.
• Event batching where vendors support it.
• Concurrency settings.

11. Common mistakes

Part of the GTM Server-Side series · RGM Training