Why agents get logged out with a LAGGED message and how to fix it

An agent is mid-shift, the screen freezes for a beat, and then it kicks them with a LAGGED warning. They log back in, it happens again an hour later. It feels random, but it isn't — LAGGED is the dialer telling you the agent screen stopped checking in on time.

What LAGGED actually means

The agent screen sends a small, steady heartbeat to the dialer — a Keepalive poll that says "I'm still here, what's next." When several of those polls arrive late or not at all, the dialer decides the Agent session is stale and drops it rather than route a live call into a screen that may be gone. That is the LAGGED logout. It is almost never the agent doing something wrong; it's the connection or the box falling behind.

flowchart TD
  A[Agent screen sends heartbeat] --> B{Arrives on time}
  B -->|Yes| C[Session stays active]
  B -->|No| D[Polls pile up late]
  D --> E{Dialer patience exceeded}
  E -->|Yes| F[Mark agent LAGGED]
  F --> G[Force logout]
  E -->|No| C

Confirm it in the reports

Don't guess from one complaint. Two reports turn an anecdote into a pattern.

1. Open the Agent LAGGED Report for the day in question. The first table lists the server IP of the dialer that experienced the lag and how many times it happened — that tells you if one box is the source.
2. Read the second table. It breaks each event down by agent log ID, user ID, server IP, event time, lead ID, campaign, status, user group, dial method, and unique ID. Look at who and when — one agent or everyone, one moment or all shift.
3. Open the LAGGED Summary Report. Its timeline chart bunches LAGGED events into clusters, so a spike at, say, 2pm every day jumps out where the raw rows hide it.

What to change

The reports point at one of two causes. If a single agent or a single remote site keeps showing up, it's their connection — the heartbeat is racing across a flaky link. If everyone clusters at the same time on the same server IP, the box is overloaded and can't service the polls.

• One agent or one site: check their network. A Remote agent on shaky home Wi-Fi or a saturated office uplink will lag where a wired desk won't. Rising Latency on that path is the smoking gun.
• Everyone at once: look at server load at that timestamp — CPU, disk, and the number of agents the box is asked to carry. A dialer pinned at peak can't answer heartbeats fast enough.

Fix the network for the one-agent case, shed load or move agents off a hot box for the everyone case, then re-pull the LAGGED Summary the next day to confirm the cluster shrank. For the full symptom-to-report map, see the troubleshooting playbook, and to read the LAGGED tables column by column, see how to read the Agent LAGGED Report. If you'd rather not size and babysit the dialer yourself, VICIfast runs a right-sized managed box so server-load LAGGED stays off your plate.