Diagnosing LAGGED spikes by server with the summary chart

Agents say the screen froze. Calls seem to stall. Then it clears on its own. This is usually a LAGGED event, and chasing it call by call gets you nowhere. The fast way to understand it is to step back and look at when the lag happened and which box it happened on. Two reports do exactly that, and together they turn a vague "the screen lags" complaint into a precise picture of where the trouble lives.

What LAGGED actually means

An agent screen stays in step with the dialer by checking in constantly over the web. When that loop falls behind, the agent's screen has lost sync with the dialer — the dispositions, timers, and call state on screen no longer match what the box thinks is happening. VICIdial records that as a LAGGED event. A handful across a day is normal. A burst in a short window is a sign that something on the server or its network hiccuped, and that is what you are hunting for.

Start with the Agent LAGGED Report

This report has two halves. The first half lists, per server IP, how many times lag occurred on that box in your date range. That alone often answers the question — if one server's count dwarfs the others, your problem has a location. The second half is the detail: each lagged event with its log ID, the user, the server IP, the event time, the lead ID, the campaign, the status, the user group, and the unique ID. Use the first half to find the guilty server, then the second half to see exactly which agents and moments it hit.

The field-by-field tour is in how to read the Agent LAGGED Report.

Spot the cluster on the summary chart

A list of events is hard to read for timing. The LAGGED Summary Report plots those same events on a timeline over your date range, so clusters jump out as spikes on the chart. A flat scatter of lag across the day is background noise. A tall spike at 2:15 on Tuesday is an incident. That shape is the whole point — it tells you whether you are dealing with a steady low-grade issue or a specific event to investigate.

Reading the chart in detail is covered in how to read the LAGGED Summary Report.

Confirm with the latency reports

LAGGED tells you the loop fell behind. The latency reports tell you by how much and when. The Agent Latency Report charts each agent's web connection Latency across the day, so you can see the latency climb right before the lag spike. The Latency Gaps Report shows missing segments of latency logging while an agent was supposed to be logged in — those gaps are often the exact moments a screen session struggled. Together they confirm whether the spike on the summary chart was a network slowdown or a true outage in the agent connection.

• Agent Latency Report: did latency rise before the spike? See how to read the Agent Latency Report.
• Latency Gaps Report: were there missing log segments at the spike time? See how to read the Latency Gaps Report.

The diagnosis path

flowchart TD
  A[Agents report frozen screens] --> B[Agent LAGGED Report]
  B --> C[Count lag per server IP]
  C --> D[LAGGED Summary chart]
  D --> E{Spike on one server?}
  E -- Yes --> F[Inspect that box at that time]
  E -- No --> G[Suspect shared network or DB]
  F --> H[Agent Latency and Latency Gaps]
  G --> H
  H --> I[Cause confirmed]

A spike on one box is good news in a way — it means you have one place to look, not a whole cluster. Check that server for the usual stalls: a Real-time report that was slow at the time, a heavy report running, a backup, or the dialer's own Keepalive processes struggling. A lag pattern that follows the network rather than a box sends you to your connectivity and database layer instead. The wider method is in the VICIdial troubleshooting playbook.

VICIfast runs a single-tenant, hardened VICIdial box so your agent screens are not fighting other tenants for the same server, with health monitoring built in and live in under 40 seconds. See our plans and pricing.