Overview
Describes how Cisco Catalyst SD-WAN automatically suspends unstable BFD sessions to prevent repeated flapping and reduce unnecessary traffic disruption.
A BFD session flap is a network condition that
-
occurs when a BFD session repeatedly transitions between up and down states,
-
happens because one device in the session becomes unavailable and then available again, or
-
repeatedly recovers and fails due to unstable connections, disrupting applications and causing unnecessary traffic steering between overlay paths.
Automatically suspending BFD sessions
To prevent repeated BFD session flaps, Cisco Catalyst SD-WAN automatically suspends unstable BFD sessions based on the following parameters:
A flap cycle includes this sequence:
-
The BFD session is in the up state.
-
The BFD session transitions to the down state.
-
The BFD session comes back up.
An SLA threshold determines when to add a BFD session to the suspended list. It defines a limit for traffic metrics such as loss, latency, or jitter. When any metric exceeds the defined threshold, the system suspends the BFD session. These thresholds represent the traffic performance levels specified in the SLA.
An SLA threshold is optional. If you configure one, set higher values for loss, latency, and jitter to prevent conflicts with SLA parameters defined in SLA classes. For more details on SLA classes, see the Cisco Catalyst SD-WAN Policies Configuration Guide.
Benefits of automatically suspending BFD sessions
-
You can manually remove the affected circuit or tunnel interface from the BFD suspended list.
-
Provides monitoring of a suspended tunnel.