Resilience is layered: a site survives a bad circuit, a dead carrier, and — where you deploy it — a failed edge appliance, without users noticing. Most of that comes from sub-second failover across bonded links; hardware and carrier diversity cover the rest.
Link failover
The edge probes every transport continuously. When a link crosses a loss/latency threshold or goes down, in-flight sessions move to a healthy link in under a second, so calls and app sessions stay up.
| Setting | What it does |
|---|---|
| Health thresholds | The loss/latency/jitter levels that mark a link "degraded." |
| Failover order | Which backup takes over, and in what order. |
| Hold-down / flap damping | Prevents rapid flip-flopping when a link is unstable. |
| Failback | Whether to return to the primary automatically once it recovers. |
Carrier & circuit diversity
Because the edge bonds multiple transports, no single provider can take a site down. Combine a wired primary with a wireless (LTE/5G) backup on a different carrier for true path diversity — the most common cause of an outage is one carrier, not the internet.
Edge redundancy
For sites that can't tolerate any downtime, deploy two edge appliances in a high-availability pair. If the active edge fails, the standby takes over the site's traffic; management and policy stay identical because both pull from the same central configuration.
Test failover safely
Schedule a test window
Pick a low-traffic window and note the site in change management.
Disable the primary transport
From the Network Portal, administratively down the primary link (don't unplug it) and watch traffic move to the backup in the live telemetry.
Verify sessions held
Confirm a voice call or app session stayed connected through the switch, then re-enable the primary and confirm failback.
Match backup capacity to critical traffic
A backup link doesn't need to match the primary's full bandwidth, but it should carry your real-time and business-critical traffic comfortably. Use QoS so that, on a smaller backup, voice and critical apps are protected while bulk traffic is throttled.