How a Phone Alias Keeps Agents Online When a Dialer Fails

A dialer failing in the middle of a shift used to mean scrambling to re-credential a chunk of your floor. With phone aliases it is a one-minute recovery: the affected agents log out, log back in with the same alias, and the system puts them on a server that is still up. Nothing about their login changes.

What happens when a server drops

A server leaves the rotation in two ways: it crashes outright, or an admin flips it to Active=N to drain it. Either way it stops appearing in the list of active dialers that the alias balances across. The agents who were on it are now on a dead line - but every other phone behind their alias is still registered and waiting.

Because the Softphone holds a parallel registration to each dialer, the healthy lines never went anywhere. The only thing that needs to refresh is the agent's session, and that happens on the next login.

stateDiagram-v2
  [*] --> OnDialerA
  OnDialerA --> DialerADown: server fails
  DialerADown --> LoggedOut: agent logs out
  LoggedOut --> Rebalancing: agent logs in with alias
  Rebalancing --> OnDialerB: only healthy servers counted
  OnDialerB --> [*]

The recovery steps

1. The agent logs out of the agent screen.
2. They log back in with the same alias, phone password, and user login.
3. The load-balancer counts only active servers and routes the agent to a healthy dialer.

This is real Failover without a separate high-availability layer. It does not move a live call that was already up when the server died - that call is lost like any dropped connection - but it gets the Agent back to taking calls in seconds, and it does so without a Keepalive race or any manual re-pointing of the line.

For the registration that makes this recovery instant, see how to test phone registration, and the phones and aliases guide for the full setup.

VICIfast runs a redundant cluster so a single box failing is a non-event for your floor - see pricing.