Study first
Review the ideas behind the questions
Review the production details that decide whether server-side GTM can support live measurement safely. Focus on capacity, preview routing, cluster consistency, logging costs, health checks, and update habits.
Plan live traffic before relying on the server
A working server-side GTM endpoint is not the same as a production-ready deployment. Capacity, redundancy, domain setup, logging, and health checks all affect whether live measurement can be trusted.
- The default automatic server-side GTM deployment is recommended for testing limited traffic, not as the final plan for heavy live traffic.
- Cloud Run capacity planning should account for instance range and tag workload, because performance varies with tag count and what tags do.
- High request volume can create logging costs, so request logging should be part of production cost review.
In Practice
Capacity is not only traffic
Use request volume as a first estimate, then adjust for how many server-side GTM tags run and what those tags do.
Health checks are part of operations
The server URL plus the health path can verify that the server responds, which helps separate infrastructure failure from tag-logic issues.
Common mistakes
Waiting for reports to reveal infrastructure capacity problems.
Plan redundancy and Cloud Run capacity before live measurement relies on server-side GTM.
Q&A
Is the default server-side GTM deployment a final high-traffic plan?
No. It is recommended for testing limited traffic, so live traffic needs additional capacity planning.
What should a health check return for a working server container URL?
The health path should return ok when the service works.
Keep preview and cluster infrastructure exact
Enterprise preview routing and manual clusters have specific infrastructure rules. Small differences in preview URL, routing style, timeout, or cluster settings can make review evidence unreliable.
- A private preview server should use host-based routing through a load balancer with a separate preview subdomain.
- Path-based preview routing is not recommended because debug can open while events fail to reach the correct destination.
- Manual preview and tagging servers use separate settings: the preview server runs as preview, while the tagging cluster points to the preview server URL.
In Practice
Preview server is special
Do not scale a manual preview server like a production cluster. The manual guide says the preview server should be exactly one instance.
Timeouts can break preview
If a CDN or load balancer in front of preview has too short a timeout, preview can fail even when the server exists.
Common mistakes
Assuming a debug window means private preview routing is correct.
Check that preview events also appear; path-based routing can open debug while losing forwarded events.
Q&A
What routing style should a private preview server use?
Use host-based routing with a separate preview subdomain.
Should every tagging server in a manual cluster use matching settings?
Yes. Each tagging server instance should use the same container configuration and preview server URL settings.
Maintain the server after launch
Server-side GTM operations continue after the first release. Owners should verify custom URLs, health responses, update messages, and logging behavior before small infrastructure issues become measurement gaps.
- After setting up a tagging server, add the server URL in the GTM server container settings.
- If multiple server URLs are mapped to one tagging server, their URL paths must match for preview selection.
- Server updates include security vulnerability fixes and new features, so owners should update when GTM reports the need.
In Practice
Check URLs before preview
When several domains point to one tagging server, mismatched paths can make preview selection fail in a way that looks like a GTM issue.
Verify updates in preview
After deploying an updated tagging server, use preview and check the console so the team knows the update message is gone or understands any remaining lag.
Common mistakes
Treating server updates as normal GTM workspace edits.
Deploy the updated tagging server revision and verify it in preview; workspace publishes do not update the server image.
Q&A
What should the owner do after changing a server-side tagging server URL?
Add the server URL in the GTM server container settings and validate it.
How should a server update be checked?
Start preview, send a request, and check the console for update messages.