5 Ways a FortiGate HA Failover Gets Triggered — Explained

Your FortiGate HA cluster just failed over. But why? A FortiGate HA failover can be triggered by different conditions, and knowing which one occurred is essential for troubleshooting. In Episode 8 of the FortiDepth FortiGate HA series, we explain five practical failover categories: device failure, monitored interface failure, remote link monitoring, memory-based failover, and SSD failure. You'll learn what triggers each category, which features require configuration, how to investigate a failover using HA history, and why an unmonitored WAN interface can leave your cluster unable to respond to an upstream failure. The episode finishes with a live HA failover demonstration, a Wireshark capture of heartbeat traffic, and a look at the gratuitous ARP associated with the transition. What you'll learn Device failover when the Primary loses power or becomes unavailable Link failover when a monitored interface goes down Remote IP monitoring for detecting failures beyond the directly connected network Memory-based failover and its configurable thresholds SSD failover and the conditions that can trigger it Which failover mechanisms require explicit configuration How to identify a failover event using diagnose sys ha history read Why unmonitored interfaces can leave an upstream failure undetected How heartbeat traffic changes during a failover How gratuitous ARP helps notify the network of the new Primary Chapters 00:00 What Just Failed Over? 00:30 Five Ways a Failover Starts 01:30 Device Failover 03:00 Link Failover 04:30 Remote Link Failover 06:00 Memory-Based and SSD Failover 07:15 How to Identify the Trigger 08:15 The Common Mistake — Unmonitored WAN 09:15 Live Packet Capture — Heartbeat and Failover 11:15 Recap and Episode 9 Preview The five failover categories *1. Device failover* The Primary becomes unavailable, for example because it loses power or stops communicating over the HA heartbeat. *2. Link failover* A monitored interface loses link. HA can react to the failure only if the relevant interface is included in the monitoring configuration. *3. Remote link failover* Remote IP monitoring detects failures beyond the directly connected network, such as an unreachable upstream gateway or another monitored target. *4. Memory-based failover* When enabled, this feature can trigger a failover if memory utilisation remains above the configured threshold for the required monitoring period. *5. SSD failover* When enabled and supported by the hardware, SSD failure detection can trigger an HA failover. How do you find out what happened? Use diagnose sys ha history read to investigate historical HA events and identify the recorded reason for a state change. Correlate the event with system logs, interface state and heartbeat observations rather than relying on a single indicator. FortiGate HA Series Episode 1 — Build Your First FortiGate HA Cluster Episode 2 — FortiGate HA Architecture Explained Episode 3 — How a Primary Gets Elected Episode 4 — Virtual MAC Addresses and Why Failover Is Seamless Episode 5 — Configuring HA Synchronisation Episode 6 — Verifying an HA Cluster Is Actually Healthy Episode 7 — Reserved HA Management vs In-Band Management Episode 8 — HA Failover Types Explained Next episode: Session Pickup — what survives a failover, tested protocol by protocol. Subscribe to FortiDepth for practical FortiGate HA labs, real packet captures, CLI troubleshooting and network security tutorials. Tested on FortiOS 7.6.7. #FortiGate #Fortinet #FortiGateHA