It sounds like you've run into a common frustration with integrating Google Analytics directly into ArcGIS Enterprise Sites.
The core issue you're experiencing with the URL fragment (`#`) not being accepted in Google Analytics property settings is typical for Single Page Applications (SPAs), which ArcGIS Sites are.
The fragment is used by the application for client-side routing, and traditional Google Analytics setups often struggle to track changes after the hash, seeing it all as one page unless configured otherwise.
The lack of native tracking for all elements, as mentioned by Arpa_PiemonteGeoportal_Team, especially within components like the map and scene viewers, is a significant limitation of the built-in functionality.
The workaround of injecting the `gtag.js` script into the server files, as suggested by akochman, is a valid temporary fix, but as PLadd notes, it is prone to being overwritten during upgrades, making it difficult to maintain.
A more robust and scalable workaround involves using a combination of the ArcGIS APIs, Google Tag Manager (GTM), and a server-side solution, which gives you much finer control over what you track.
First, the ArcGIS API for Python and the underlying ArcGIS REST API can be used to programmatically monitor and gather information about your content and user interactions that happen on the server-side or are exposed via the API, such as item views or downloads, bypassing the front-end GA integration problem entirely for some metrics.
This data can be structured and sent to a data warehouse or directly to a reporting system.
Second, for front-end events and page views that the embedded script misses, the best approach is a server-side tagging solution using Google Tag Manager and a service like Stape or even a custom setup on Google Cloud Platform.
With GTM implemented, you can set up custom events to fire when users interact with the site.
The critical part is using GTM to push these interactions not just to the user's browser but to a server container hosted on Stape or GCP.
This server-side tagging setup is then configured to forward the data to Google Analytics (GA4 is recommended).
This approach addresses the SPA issue by allowing you to manually fire a virtual pageview or Standard Events like `page_view`, `view_item`, or `click` from JavaScript inside the Site pages when the URL hash changes or when a specific action occurs (like opening a map viewer).
Finally, you can bring all this data together using the Google Analytics Data API.
This allows you to programmatically access all the tracked data from your GA property, including the custom events you're sending from your server container, and combine it with the server-side metrics you gather from the ArcGIS REST API.
This method provides a comprehensive view of user behavior across both the core site structure and the richer, dynamic map and scene viewer components that the native GA integration fails to monitor.
While it requires more initial setup and maintenance, this strategy future-proofs your analytics against ArcGIS Enterprise upgrades and provides a depth of tracking not possible with the native, limited feature.
.png){kind=link}