A FortiGate can pull a plain-text address list from an external URL and turn it into an object you can use in firewall policies. Fortinet calls this a Threat Feed, or an external resource in the CLI. This guide walks through wiring the IP-DB.com pool into a FortiGate, using it in policy, verifying it, and fixing the usual failures.
Prerequisites
- A FortiGate with admin access (hardware or VM)
- Outbound internet from the device: 443/TCP and working DNS resolution
- A IP-DB.com API key, created from your account dashboard
- A correctly set system clock — certificate validation depends on it
1. Choose the feed URL
The whole pool is served from one plain-text endpoint, one CIDR per line:
To narrow the list to one threat type, append &category=. Valid values include botnet_c2, malware_c2, attack_source, ssh_brute_force, compromised_host, hijacked_network, tor_exit, and aggregated_high_confidence. Narrowing does more than cut noise — it also keeps you clear of the device's entry limit.
2. Create the external connector
In the GUI: Security Fabric → External Connectors → Create New → Threat Feed → IP Address. The CLI equivalent:
refresh-rate is in minutes. Because the pool is rebuilt every 30 minutes, 30–60 is the right range; dropping to 5 fetches no new data and simply burns the 120 requests per hour your key allows.
3. Use it in policy
The resulting object shows up in policy address fields like an ordinary address group. Use it in both directions:
- Inbound (WAN → internal): a deny rule with this object as the source, placed above your allow rules. Most scanning and brute-force traffic dies here.
- Outbound (internal → WAN): as the destination, using a second feed scoped to
botnet_c2andmalware_c2. That rule stops a machine inside your network from reaching its command-and-control server — and it is usually where you first notice an infection.
Keep logging enabled on both rules; without it you cannot measure whether the feed is doing anything.
4. Verification
The View Entries link on the connector card shows how many entries the device actually loaded. From the CLI, configuration and status look like this:
A count of zero means the list never downloaded; a count lower than expected usually means you hit your model's entry limit — narrow the feed by category.
5. Troubleshooting
- Connector shows red: start with
execute ping ip-db.comand your DNS settings. No name resolution, no feed. - Certificate error: validation fails when the clock has drifted. Check
execute timeand your NTP configuration. - Empty list: confirm the API key was copied into the URL in full; with an invalid key the endpoint returns an authorization error.
- Intermittent failures: the same key across many devices every 15 minutes will exceed the hourly limit. Issue a key per device, or widen the interval.
Frequently asked questions
Which FortiOS versions support threat feeds?
The external-resource object exists on all current FortiOS releases; from 6.4 onwards it appears in the GUI under Security Fabric → External Connectors. On older builds you configure the same object from the CLI only.
How many addresses can the feed carry?
The external resource entry limit depends on the model, from a few thousand on small desktop units to tens of thousands on larger platforms. Scope the feed with &category= — botnet_c2 and attack_source cover most cases.
What refresh interval should I use?
The pool is rebuilt every 30 minutes, so keep refresh-rate between 30 and 60. Polling more often returns no new data and burns through the 120 requests per hour allowed for your API key.
What if the connector stays red?
Check, in order: DNS resolution on the FortiGate, outbound access to port 443, the API key, the system clock (certificate validation fails on clock skew), and whether you have exceeded the hourly request limit.
Short setup recipes for other platforms live on the integrations page, and endpoint details in the API documentation. For how long and how wide to ban, see our blocking guide.
Comments 0
Share a question, your experience or a correction. You can comment as a registered user or anonymously.
No comments yet. Be the first.