LLT interconnect link monitoring and alerting
On multi-node appliances, the nodes communicate with each other over private LLT (Low Latency Transport) interconnect links. These links carry the cluster heartbeat traffic between the nodes. If enough of these links go down at the same time, the cluster can enter a degraded state in which a single additional failure could split the cluster.
Flex Appliance monitors these links and alerts you early, without generating false alerts during planned maintenance, so you can resolve interconnect problems before they affect availability.
Flex Appliance monitors the following LLT links on each node of a multi-node appliance:
privnic0
privnic1
mgmt0
Note:
This feature applies only to multi-node appliances. It does not apply to single-node appliances, which do not have LLT interconnects.
Key features:
Continuous monitoring of interconnect link status as part of the appliance's existing health monitoring cycle. No additional configuration is required to enable it.
Automated alerting when a link is reported down.
Alerts are sent through your existing email and SNMP alert configuration.
Alerts are throttled to once every 24 hours per affected node, severity level, and link combination. If the affected links change, a new alert is raised immediately. If the same combination persists, you will not receive another alert for that node and severity until the 24-hour window expires.
LLT link warnings and faults are also visible on the Flex Appliance Console dashboard, under the widget.
Note:
Each node independently monitors its own connectivity to its peer. If an LLT link breaks, each node can no longer reach its peer over that link, so both nodes raise their own alert for the warning or fault. You may see an alert from each node for what is a single physical link failure.
Flex Appliance evaluates how many of the three tracked LLT links are down on a node and raises an alert based on the following thresholds:
Table:
Severity | Trigger condition |
|---|---|
Warning | Exactly one of the three tracked LLT links is down on a node. |
Critical | Two or more of the three tracked LLT links are down on the same node. This indicates a greater risk to cluster availability. |
Both alert types include:
The affected node.
The specific link or links that are down.
Recommended corrective actions, such as checking cable connections, network interface link status, and switch port configuration.
Note:
Flex Appliance does not raise a link-down alert while a peer node is in a recognized offline transition, for example during a planned shutdown, an appliance update, or a rollback. In these situations, links to that peer are reported as unknown rather than down until the node returns to normal operation.