Our free automatic failover system constantly monitors the network for disruptions and instantly switches to another server if any issues arise with your selected server.

Each Cloud site has a primary and secondary origin server located in different regions that mirror one another and remain in constant sync.

In the event of a server or data center outage, our origin routers detect the failure and immediately start sending traffic to the secondary server in the site’s pool. This server is designed to serve traffic in a “read-only” state until engineers confirm the primary server has failed. At this point, using the platform API, the relationship between the primary and secondary server is severed and the new server begins serving traffic in “read-write” mode. This allows for automatic zero, or near-zero, downtime.

How Automatic Failover Works

Group Failure: This occurs when an administrator-defined group of server nodes within a cluster becomes unresponsive due to power loss or a network issue affecting a specific subnet or rack. All customer data is continuously replicated to an alternate region, enabling automated region-based failover. If an entire region goes down, traffic is seamlessly redirected to another region.

Disk Read/Write failure: When a disk on a specific node experiences repeated read/write errors over a prolonged period, the system triggers automatic failover. While the node remains reachable, it is removed from active service to prevent disruptions.

Node Failure: A node in the server cluster may become unresponsive due to out-of-memory errors, network failures, or other node-specific issues. The cluster manager automatically detects and verifies the problem, triggering a hard failover. This ensures your site remains online by switching traffic to a backup server. Once the issue is resolved by an administrator, your site seamlessly returns to the original node.

Graceful Failover

Unrelated to data center and server outage failovers, individual sites that leverage Edge Cache and encounter plugin, theme, or code snippet errors may benefit from Graceful Failover. This is where cached versions of a page are automatically served if the origin returns a 5xx HTTP status response.