Service eligibility and regional restrictions
PuppyIP serves only compliant overseas businesses and their authorized personnel. Proxy services are not available in mainland China. The service may only be used for lawful business activities outside mainland China. Use of this service within mainland China is prohibited.
Hosting a proxy IP or server overseas does not change these restrictions. The service must not be provided to end users in mainland China through relaying, forwarding, sharing or resale. Before use, read the Terms of Service.
Key Takeaways
- The November 17, 2026 update adds an optional Secondary gateway address for static WANs on HA sites only.
- Leaving it blank retains the primary gateway. Separate gateways let the appliances use their own static addresses and Internet circuits, including different ISPs.
- The Cloudflare WAN overview labels the product Enterprise-only. Appliance costs, contract entitlements and regional supply require separate confirmation.
- Primary/Secondary identifies the appliances; Active/Standby describes their current roles. Fully restoring traffic after failover may take up to 30 seconds.
An HA pair can now use separate WAN gateways
On November 17, 2026, Cloudflare announced that each Cloudflare One Appliance in a high availability, or HA, site can have its own static WAN default gateway. WAN is the side connected to the external network. Previously, both appliances shared one gateway; they can now connect to their own Internet circuits, such as a different ISP for each appliance.
The new field applies to sites already in HA mode and to WANs with static addressing. Do not assume it applies to ordinary single-appliance sites, DHCP WANs or every connection type. An existing HA site that still uses one shared gateway does not need a configuration change simply because the field exists.
Cloudflare WAN is an enterprise wide-area networking service, and its official overview explicitly labels it Enterprise-only. The reviewed sources do not establish a universal price for this HA update, appliance purchase costs, country-by-country supply or each account’s contract entitlements. Confirm access with the business responsible for the deployment; the announcement does not establish free access or automatic availability for every account.
Match each WAN address to its own gateway
In a static WAN configuration, Static address belongs to the primary appliance and Secondary static address to the secondary appliance. Primary gateway address is the former’s default gateway. The new Secondary gateway address is the latter’s default gateway. Different interface addresses alone do not establish that the appliances use separate gateways.
If the secondary appliance’s circuit has its own gateway, enter it in Secondary gateway address for the relevant static WAN on the existing HA site. Leaving it blank continues to use Primary gateway address and preserves the previous behavior. Match each interface address, prefix length and gateway to the values actually assigned to that circuit.
The fragment below uses the announcement’s static_addressing structure and reserved documentation addresses to illustrate two circuits with separate addresses and gateways. It explains field mapping; it is not a complete API request or addressing suitable for a real network. Preserve the WAN’s other existing settings.
Static WAN fields for separate gateways
json{
"static_addressing": {
"address": "203.0.113.2/30",
"gateway_address": "203.0.113.1",
"secondary_address": "198.51.100.2/30",
"secondary_gateway_address": "198.51.100.1"
}
}
Handle existing HA sites and new HA sites separately
For an existing HA site, check the optional gateway field on the target WAN. Do not recreate the site just to use separate gateways. Preserve its current configuration and circuit parameters, have an authorized network administrator plan the change and recovery method, and inspect the saved configuration.
A new HA site requires two appliances deployed on your premises first. After naming the site in the dashboard, select Turn on high availability and then Create and continue. Use Add Appliance for the first device and Add secondary appliance for the second, confirming that both belong to the intended site.
Next, create a WAN. For static addressing, enter the primary and secondary interface addresses separately. If their circuits use different default gateways, also enter Secondary gateway address. Configure the LAN, select High availability probing link, and follow the official appliance setup and activation steps.
An ordinary existing site cannot have HA enabled in place. The dashboard guide requires deleting the site and starting again. That is a migration that may affect connectivity, not a routine troubleshooting toggle. For a site carrying traffic, the network owner should first approve the migration plan, maintenance window, preserved configuration and recovery path; do not delete the site as an unplanned step.
Use the LAN virtual gateway and a probing link through a switch
LAN is the side connected to the local network. Each LAN needs three static addresses: the primary appliance’s address, the secondary appliance’s address and Virtual static address. The virtual address is the gateway used by downstream LAN devices. All three LAN addresses must be in the same subnet. Do not substitute a WAN default gateway for the LAN virtual address.
Choose a reliable interface under High availability probing link. The appliances exchange heartbeats over this LAN link to elect the active node. Cloudflare requires the probing link to run through a switch; the appliances must not be connected back-to-back.
With a direct connection, unplugging one end also makes the other end detect link failure. A failed HA link puts the affected node into FAULT, preventing it from acting as the active node. Changing a WAN gateway does not fix that probing-link topology problem.
Which appliance forwards traffic, and how long can recovery take?
Primary/Secondary is the identity assigned in the dashboard to distinguish appliance configuration. Active/Standby describes the roles selected dynamically. Only one appliance is expected to be active at a time. A recovered primary appliance is not necessarily forwarding traffic again, and an HA pair does not mean that both appliances simultaneously forward all business traffic.
The higher-priority appliance becomes active; when priorities match, the state machine selects one. HA uses non-preemptive operation: a node that becomes active keeps that role unless its priority falls below the other node’s.
The reference distinguishes GOOD, DEGRADED and DOWN using tunnel, LAN-link and software health. Failover occurs when the active node’s health falls below the standby node’s. Traffic may take up to 30 seconds to be fully restored on the new active node. This describes documented behavior, not lossless recovery for every application session, a fixed recovery time or a business SLA.
Verify configuration, roles and actual business traffic separately
Read back both WAN address-and-gateway mappings and match them to the physical circuits. Then check the subnet of the three LAN addresses, downstream devices’ default gateway, and the probing connection through a switch. A successful save establishes that settings were accepted; it does not establish that traffic followed the intended circuit.
An authorized network administrator should inspect appliance roles, tunnel health and required business paths within an approved maintenance arrangement. If testing circuit or appliance failure, define the expected active node, acceptable interruption and stop conditions first. Observe the resulting role and business recovery and record the outcome; do not unplug production circuits casually for a test.
If traffic fails after the secondary appliance takes over, check its Secondary static address and Secondary gateway address against its actual circuit, followed by the LAN virtual gateway, probing link and tunnel health. This article verifies public documentation only. No real appliance configuration or failover test was performed, and no recovery time was measured on a business network.
Sources
Frequently Asked Questions
How do I clear the secondary appliance’s separate gateway through the API?
The announcement names static_addressing.secondary_gateway_address. Set it when creating or updating a WAN; to remove it with PATCH, set that field to null. Removing it makes the secondary appliance use the primary gateway again, so first confirm that its circuit can still use that gateway. No API request was executed for this article.
Do I need to create IPsec tunnels and static routes manually after enabling HA?
The Cloudflare One Appliance configuration guide says the appliance creates IPsec tunnels and static routes automatically and that these objects cannot be configured manually. Do not apply general IPsec/GRE tunnel BGP setup steps directly to objects generated by the appliance.
Does appliance failure automatically bypass the device and allow traffic through?
The official reference explicitly states that Cloudflare One Appliance does not support fail open. A standby node taking over in HA and traffic bypassing a failed appliance are different mechanisms. Support for the former does not establish support for the latter.