📘 GA Data to Azure Solution Option: GA4 → Server GTM → Azure Databricks

Architecture Overview

The following diagram illustrates the end-to-end data flow:

flowchart LR U["User
Browser"] W["GTM Web Container
Fires GA4 Config Tag"] S["sGTM Server Container
Receives events via transport_url"] GA["Google Analytics
(Reporting)"] AF["Azure Function
HTTP Trigger"] EH["Azure Event Hub
(Capture Enabled)"] ADLS["ADLS Gen2
Raw Event Storage"] DB["Azure Databricks
Autoloader → Delta Tables"] U --> W W --> S S --> GA S --> AF AF --> EH EH --> ADLS ADLS --> DB

Data Flow: Browser → analytics.beyondscreen.online (sGTM) → GA4 + Azure Function → Event Hub → ADLS → Databricks

1. Website Setup

Host website on GitHub Pages (prasoot.github.io).
Register custom domain in GoDaddy (beyondscreen.online).
Configure GoDaddy DNS:
- www.beyondscreen.online → CNAME → prasoot.github.io
- Optional root ANAME/ALIAS → GitHub’s IPs.
Add GTM container snippet to <head> and <body>.
Remove direct gtag.js GA snippet (avoid duplicate hits).

2. GTM Web Container

Create GA4 Configuration Tag (Measurement ID: G-L70NLL6EC4).
Set transport_url = https://analytics.beyondscreen.online.
Create GA4 Event Tags linked to this Config Tag.
Publish and verify in DevTools → network → collect.

3. First Party subdomain setup for sGTM

Create subdomain: analytics.beyondscreen.online → CNAME ghs.googlehosted.com.
Map custom domain in Cloud Run and wait for SSL.
Verify health check: https://analytics.beyondscreen.online/__gtmhealthz → 200 OK.

4. sGTM Server Container (Cloud Run)

Deploy server container. Add GA4 (Web) client.
Add GA4 Forwarder Tag and Azure Forwarder Tag (HTTP POST to Azure Function URL).

5. Azure Function

Create Function App → Python 3.10 → Consumption plan.
Add HTTP Trigger function (ingest).
Code parses GA4 JSON and forwards to Event Hub.

6. Azure Event Hub

Create Event Hub namespace (ga4-hub-ns).
Create Event Hub (ga4-events).
Enable Capture → ADLS Gen2 (Avro/Parquet).
Verify metrics → “Incoming Messages” increasing.

7. ADLS Gen2

Create storage account with hierarchical namespace.
Create container ga4-raw.
Event Hub Capture writes Avro/Parquet files here.

8. Azure Databricks

Create workspace and mount ADLS Gen2.
Ingest via Autoloader:

9. Validation

Browser DevTools → only analytics.beyondscreen.online.
sGTM Preview → GA4 + Azure forwarders firing.
Azure Function logs → GA4 ingest function triggered.
Event Hub metrics increasing.
ADLS → new files arriving.
Databricks → row counts increasing.

✅ End State

All GA4 hits go via your first-party domain → sGTM → GA4 + Azure → Event Hub → ADLS → Databricks.
You get analytics in GA + raw data in Azure for advanced analysis & ML.