Cannot reach the FortiGate GUI, SSH, or ping? Follow a focused checklist for interface access, trusted hosts, local-in rules, and the return path.
Client through firewall to server
Client to a firewall service
Firewall response back to client
First identify the destination
A firewall can pass internet traffic successfully while its own management address remains unreachable. These are different traffic paths. A forward firewall policy controls traffic crossing the device. Local-in controls apply to traffic terminating on the FortiGate itself. Adding another LAN-to-WAN allow rule does not explain why a browser cannot open the firewall GUI.
Write down the client address, exact destination interface address, protocol, port, ingress interface, and VDOM. Use the configured HTTPS port rather than assuming TCP 443. If the address is a VIP for a server behind the firewall, investigate the forwarded connection instead.
Check the interface and administrative settings
Review the target interface configuration and the system settings in the correct administrative context. Confirm that the intended management protocol is enabled on the relevant interface. Then review the administrator account’s trusted-host restrictions and authentication requirements. These checks answer different questions and should be recorded separately.
Use the following read-only commands as starting points. Output may contain infrastructure details, so redact it before sharing. Exact options vary by FortiOS release and deployment mode. These commands are documentation-based examples, not a claim of a test on your appliance.
show system interface
show system global
show firewall local-in-policy
Review local-in rule scope and order
Inspect source objects, destination objects, incoming interface, service definitions, schedules, action, and sequence. A source object must match the address that actually reaches the firewall after any upstream translation. A rule that looks correct for a private laptop address may miss when the laptop is behind an internet NAT gateway.
Do not assume an absent user-created local-in rule means that every service is exposed. Administrative access settings and other platform controls still matter. Equally, an enabled HTTPS checkbox does not prove that all upstream and local restrictions permit the connection. Check IPv6 separately when the browser resolves an IPv6 address.
Capture one controlled connection attempt
Run a bounded capture while an approved client attempts one connection. Replace the example address and port. Stop after the relevant packets have appeared. An arrival capture proves delivery to an observation point, not successful authentication or GUI availability.
If no request arrives, examine the client route, VPN, upstream ACL, destination address, and NAT. If requests arrive without replies, focus on the local service, restrictions, and return route. If a TCP handshake completes but the browser shows a certificate or login error, investigate that application-layer evidence rather than adding broader network access.
diagnose sniffer packet any 'host 192.0.2.25 and tcp port 443' 4 20 l
Validate the correction without losing access
Prepare console or approved out-of-band access before changing management restrictions. Export the relevant configuration, make one scoped change, and test from both an authorized management source and an unauthorized test source. Confirm that legitimate management still works and that the restriction remains effective.
Record the ingress interface, matching control, observed reply path, and final test result. Keep this evidence with the change record. If the real problem is routed access to a server, continue with the debug-flow guide below rather than treating every timeout as a local-in failure.
References and further reading
Fortinet local-in policy guide