Pick topology, set RPO/RTO targets, and get a complete HA plan: infrastructure list, failover trigger conditions, activation steps, failback procedure, and a pre-disaster rehearsal checklist.
UCCX HA is well-documented by Cisco but misunderstood in practice. Many sites "have HA" because they bought two servers — but never tested failover, never documented the sequence, never trained ops. When the primary dies, they discover the secondary doesn't have current CUCM integration, or the engine hasn't synced in 3 weeks. This tool emits the full HA plan including the rehearsal schedule.
Pick topology + RTO/RPO, click Generate HA Plan.
What you need to buy / build. Servers, licenses, storage, network.
What triggers it, who authorizes, what runs.
How to return primary → secondary sync after recovery.
Quarterly simulation drill. Must happen.
UCCX HA is a two-node active/standby cluster. Only one node serves calls at a time — unlike CUCM where subscribers share load actively. Failover is automatic: the standby detects heartbeat loss and begins accepting calls within ~90 seconds. Calls in progress at failover drop — this is unavoidable. Queued calls are lost (the standby doesn't have queue state from the primary).
For geo-HA, WAN round-trip must be under 80ms for synchronous replication of the database. Above 80ms, DB transactions will queue and the primary will slow under load. At 150ms+ you should switch to asynchronous DR (slower RPO, but stable).
Rehearsal cadence: quarterly. Every 90 days, announce a 2-hour maintenance window, fail over to standby, run for 30 minutes, fail back. Document time to recovery actual vs target. Sites that skip rehearsals discover their HA is broken only during a real outage.
No. UCCX is strictly active/standby. Only one engine processes calls at a time.
No. Seat licenses are cluster-wide. You buy once, use on either node.
~60-90 seconds after primary engine dies. Agents need to re-login on standby.
Yes — under 50 agents with low SLA criticality, single-node is viable. Add VMware HA at the hypervisor level for resilience.
40 agents, same datacenter, 5ms WAN.
200 agents, 2 datacenters, 30ms WAN, 500Mbps.
60 agents, on-prem primary, cloud secondary, 60-min RPO.
Our UC Modernization Sprint includes HA validation + rehearsal drill + documented runbook.