An inbound rule opens one port on one protocol and carries the settings Prism
applies to traffic arriving there. Everything you tune per port, the rate
ceilings, the source restrictions, and the protection engines, hangs off a rule.
Servicesyour serverFirewallInbound rules
game-01 / FirewallExample view
Inbound rulesSource IPsProtectionBackdoors
Inbound rulesAn allow-list: only ports with a rule are reachable. Cap a port’s packet rate, restrict it to proxied sources, or drop UDP.
1Add rule
Default for unlisted portsAllow: unlisted ports pass through to always-on mitigation.AllowDeny all
Service
Protocol
Ports
Policy
Status
FiveMGame servers
UDP
30120
No extra limits
Enabled
FiveMGame servers
TCP
30120
No extra limits
Enabled
SSHInfrastructure
TCP
22
No extra limits
Enabled
FiveMGame servers
UDP30120EnabledNo extra limits
FiveMGame servers
TCP30120EnabledNo extra limits
SSHInfrastructure
TCP22EnabledNo extra limits
Quick setup: apply a preset for a common workload
The Inbound rules tab. The Default for unlisted ports sits above the rules, each rule is one row with its protocol, port and policy, and the pencil, copy and bin icons on a row edit, duplicate and remove it. Step 1 of adding a rule is the Add rule button. Example view.
Inbound rules are an allow-list. A rule names the port, the protocol, and the
service you actually run there, and that is what tells Prism how to judge the
traffic: which protocol validator to run, which packet-rate ceilings apply, and
which sources are acceptable.
A port with no rule of its own is not unprotected, but it is not tuned either.
It falls to the Default for unlisted ports setting on the same tab: under
Allow it stays reachable and is filtered only by the always-on protections,
and under Deny all it is dropped. See
The Firewall tab.
What a packet meets on its way to a port. With no rule for its protocol and port, the Default for unlisted ports decides. With a rule, it passes the rule's checks in order, and any of them can drop it; the engines run alongside, where they apply. On a new rule the switches are off, the ceilings are 0 and protocol validation is off, so the rule opens the port and those checks wait for you to set them. Illustration.
Two things about a rule's shape are worth knowing before you add one:
A rule is TCP or UDP, never both. A service that needs both, such as a
FiveM server, needs two rules.
A rule is identified by its start port and protocol together. Only one
rule can own a given pair, so there is one TCP 30120 rule and one UDP 30120
rule, not two of either.
Add port ruleOpen a port (or range). Cap its packet rate, restrict sources, drop UDP, and choose which protection engines run on it.
2ServiceGeneric / custom
3ProtocolTCP
4PortPort
5to (optional)
6Rate limits
Whole port, packets/sec (0 = unlimited)0Total accepted across every source, which is what bounds a distributed flood. Leave 0 unless you know this port’s real peak.
Per connection, packets/sec (0 = service default)0Needs a whole-port ceiling above. The datapath only consults this inside that limit, so on its own it is not applied.
7
Proxied sources onlyAccepts traffic only from known proxy or CDN source ranges.
Drop UDP on this portFor services that are TCP only. Discards every UDP packet aimed here.
Protocol validation (L7)Checks that traffic actually speaks this service's protocol and drops what doesn't.This service has no protocol validator. No effect on this one.
TCP fingerprintDrops SYNs whose TCP stack fingerprint looks forged.Turn off if a real client behind an unusual NAT, VPN or tunnel is being blocked.
Signature (zero-day)Learns the payload pattern of an in-progress attack and blocks it, without a rule being written first.Applies to UDP rules. No effect on this one.
UDP first-packet challengeUnder load, drops an unknown source's first packet. Real clients retransmit and get through; single-shot spoofed floods never do.Applies to UDP rules. No effect on this one.
Cancel
8Add rule
The Add port rule dialog as it opens. The numbers match the steps below; step 1 is the Add rule button on the tab. Example view.
Open the dialog
Press Add rule in the top right of the Inbound rules card. The dialog is
titled Add port rule and describes itself: "Open a port (or range). Cap its
packet rate, restrict sources, drop UDP, and choose which protection engines run
on it."
It opens neutral. Service sits on Generic / custom, Port is blank, and
nothing is written until you press the button at the end.
Pick the Service
Service is a grouped list: Infrastructure, Web, Database, VPN,
Voice, Game servers, and Other. Pick the service you actually run on the
port, because Prism uses it to decide which protocol checks are even possible.
An entry suffixed · L7 ships a protocol validator; see
Protocol validation.
Picking a service fills Port and Protocol in for you from that service's
usual defaults, and clears the range end. Anything you type afterwards wins.
Generic / custom has no canonical port, so it leaves the fields alone for you
to fill in.
Set the Protocol
TCP or UDP. If the service picker filled this in and your server runs the
service somewhere unusual, change it here.
Enter the Port
A whole number from 1 to 65535. A blank or invalid port is rejected with
"Enter a valid start port (1-65535)."
Set the range end, if you need one
The field labelled to (optional) covers a contiguous range. Leave it blank
for a single port. Filled in, it must be a whole number, no lower than Port,
and no higher than 65535, or the panel refuses with "End port must be ≥ start
port and ≤ 65535."
Set the rate limits, if you know your traffic
The Rate limits group holds Whole port, packets/sec (0 = unlimited) and
Per connection, packets/sec (0 = service default). Both start at 0. They
depend on each other and the numbers matter, so they have a page of their own:
Rate limits. Leaving both at 0 is a valid
rule.
Choose the switches
Below the rate limits are Proxied sources only and Drop UDP on this
port, described in the next section, then Protocol validation (L7) and the
per-rule engine switches, covered on
Protocol validation.
Press Add rule
The rule appears in the table above. It takes effect on your server at its next
configuration poll, within about 30 seconds.
Two switches on the rule narrow which packets are acceptable at all, before any
rate limit or detection engine is involved.
Switch
Field
What it does
Proxied sources only
only_proxied
"Accepts traffic only from known proxy or CDN source ranges." Ordinary players connecting directly are not in those ranges, so use it only where every legitimate client genuinely arrives via a proxy or CDN.
Drop UDP on this port
disallow_udp
"For services that are TCP only. Discards every UDP packet aimed here." On a UDP game that is every player.
Both are off on a new rule.
The same switch on two ports. Drop UDP on this port discards every UDP packet aimed at the port, whatever sent it. On a TCP-only service that is only the junk; on a UDP game it is every player, and the server looks down. Illustration: relative, not measured.
Both the browser and the server check a rule before it is written:
Field
Accepted
Port
A whole number, 1 to 65535
to (optional)
Blank, or a whole number no lower than Port and no higher than 65535
Protocol
TCP or UDP
Rate limits
Zero or above
Ownership is re-checked on the server for every change, so a rule can only ever
be written against a server on your own account.
If the start port and protocol already belong to another rule, the panel names
the clash rather than overwriting anything, with a message such as "A TCP rule
starting at 3389 already exists." Edit the existing rule instead of adding a
second one.
Plan the 15 rather than spending them. A service needing both TCP and UDP costs
two, and a contiguous range costs one rather than one per port, so use to
(optional) where the ports really are adjacent.
One server planned two ways. FiveM costs two rules because it needs TCP and UDP. Ten adjacent ports cost ten rules added one by one, or one as a range. Derived from the limits on this page; example rules.
The pencil on a rule's row opens the same dialog, titled Edit port rule,
pre-filled from the rule as it stands. The button reads Save changes.
The other trap in editing is the Service picker. Changing it rewrites
Port and Protocol from the new service's defaults and clears the range
end, exactly as it does on a new rule. If you only wanted to relabel a rule, you
have moved it: either onto a different port, or onto one another rule already
owns, which returns the collision error above. Check Port and Protocol
after changing Service, every time.
game-01 / Firewall / Edit port ruleExample view
As the rule opens
ServiceGeneric / custom
ProtocolTCP
Port8080
to (optional)8090
After picking HTTP / HTTPS · L7
ServiceHTTP / HTTPS · L7
ProtocolTCP
Port80
to (optional)
Protocol: set by the servicePort: was 8080to: cleared, was 8090
Relabelling a generic rule for TCP 8080 to 8090 as HTTP / HTTPS rewrites it to TCP 80 and clears the range end. Saving would move the rule onto TCP 80 alone and leave 8080 to 8090 without a rule, or, if a TCP 80 rule already exists, the panel refuses with “A TCP rule starting at 80 already exists.” Example view.
A saved edit reaches your server at its next configuration poll, within about
30 seconds. If a change turns out to be too strict, work through
When players cannot connect.
The bin icon on a rule's row asks you to confirm, on a dialog titled Remove
this rule? naming the rule and stating the effect: "…will no longer be
reachable. Anything connecting to it is dropped at the edge from the next config
poll." The buttons are Cancel and Remove rule.
Deleting is also the only way to change a rule's Disabled status, because
there is no control anywhere in the panel that enables or disables an individual
rule. If a rule reads Disabled, note its settings, delete it, and add it again.