Diagnosing high agent latency: the reports that point to the cause

"The screen is laggy" is the vaguest ticket you'll ever get. Buttons take a second to respond, the next lead is slow to pop, and the agent blames you while you blame their internet. Two reports settle it: the Agent Latency Report tells you how slow, and the Latency Gaps Report tells you when the screen went dark entirely.

What latency means here

Latency on the agent screen is the round-trip time for the browser's web connection back to the server — the delay between the screen asking the dialer something and getting an answer. It's separate from call audio. You can have crystal-clear RTP on the phone and a screen that crawls, because they ride different paths. High screen latency is what makes buttons feel sticky and the next-lead pop drag.

Read the Agent Latency Report first

This report shows the agent-screen web-connection latency for agents currently or recently logged in. You can pull one user's detail or every user for a single day, and it draws a chart of latency over the course of the day. That chart is the whole game — its shape tells you where to look.

flowchart TD
  A[High agent latency] --> B{Who is slow}
  B -->|One agent only| C[Their connection or Wi-Fi]
  B -->|Everyone same time| D[Server side]
  D --> E{Spike or steady}
  E -->|Spikes match report runs| F[DB or web server load]
  E -->|Steady all day| G[Box undersized]
  C --> H{Latency steady or jagged}
  H -->|Jagged| I[Packet loss or jitter on link]
  H -->|Steady high| J[Workstation or distance]

1. Pull all users for the bad day. If one agent's line sits high while the rest are flat, the problem is on that agent's side — their connection or their Agent session.
2. If every agent's latency climbs at the same time, it's server-side. Note the timestamps.
3. For a single bad agent, look at the chart's texture. Steady-high points at distance or a slow workstation; jagged spikes point at Packet loss or Jitter on their link, the same way bad audio shows up on a call.
4. For server-side spikes, line them up against heavy report runs or backups — a hammered DB or web server drags everyone's screen at once.

Then check the Latency Gaps Report

A latency gap is a missing segment of latency-log entries while an agent was supposed to be logged in — the screen stopped reporting in entirely for a stretch. Gaps are one of the most common reasons an agent has trouble with their session, and they show up as holes the Agent Latency Report alone won't make obvious. The Gaps Report charts them for all agents over the day, so a recurring 30-second hole on one user is easy to spot.

Put the two together: steady high latency on one user is their link or machine, synchronized spikes are the DB or web server, and recurring gaps are a dropping web connection that needs the network fixed. For the rest of the symptom map, see the troubleshooting playbook, and to read the gap chart in detail see how to read the Latency Gaps Report. If undersized servers keep showing up as everyone-at-once spikes, VICIfast runs a managed box sized for your agent count so the server side stays out of the latency chart.