How to read the Webserver-URL Report on a multi-server setup

You changed an agent page, redeployed, and one agent still swears they're seeing the old screen. On a single box you'd shrug and clear the cache. On a cluster with more than one web server, the real question is which server that agent actually hit — and the Webserver-URL Report exists to answer exactly that.

What this report actually shows

It is not one report — it bundles three logs, each cut by date and time range. They track web traffic to your VICIdial web servers, which is a different thing from the call logs. Think of it as the web tier's own paper trail.

• VICIdial User Log — shows where user requests are going on a cluster with multiple web servers. It logs the whole URL (in VICIdial), so you can tell which version of a page that user loaded if more than one version is live somewhere on the system.
• VICIdial API Log — the API calls received by a given server in the time range. Useful when a Non-agent API script reports against one node but not another.
• VICIdial Report Log — total reports run, per web server, for the selected window. Good for spotting one node carrying all the report load.

Why it matters on a cluster

With several web servers behind one front door, an agent's browser can land on any of them. If a deploy missed a node, that agent's Agent session is served stale code while everyone else is fine. The User Log's full URL (in VICIdial) capture is how you prove it — same path, different server, different page version.

flowchart TD
  A[Agent browser request] --> B{Load balancer}
  B --> C[Web server 1 updated]
  B --> D[Web server 2 stale code]
  C --> E[User Log records full URL]
  D --> E
  E --> F{Which server and which page version}
  F --> G[Confirm stale node]

Reading it in order

1. Set the date and time window to the moment the agent saw the wrong screen — narrow it, don't dump a whole day.
2. Open the VICIdial User Log and find that user's rows. Read the server they hit and the full URL (in VICIdial) string, including any version markers in the path.
3. Compare the same path across two users on different servers. If the URLs differ, one node is out of sync.
4. If the complaint is an integration, not a screen, switch to the API Log and confirm the call reached the node you expected.
5. Check the Report Log if reports feel slow — one server running every report is its own problem.

Once you know the offending server, the fix lives outside this report: redeploy or sync that node, then re-pull the User Log to confirm the URL (in VICIdial) now matches. For the wider map of which report answers which symptom, see the troubleshooting playbook, and for the close cousin that logs Dispo Call URLs and WEBFORM clicks, read how to read the URL Log Report. If you'd rather not run a web cluster at all, VICIfast gives you a single managed box so there's no node to fall out of sync — Managed hosting keeps the web tier to one server you never have to reconcile.