Skip to content

Bans & rate limits

IP bans

Corvus bans IPs that keep attacking, for a limited time, without you doing anything. You can also ban and unban by hand. Bans apply to all proxy and redirection hosts, not to the admin console.

IP bans

Automatic bans

Setting Default Meaning
Enable automatic bans On Ban IPs whose detections (attack patterns and body inspection, also in log-only mode, and security rules that block) exceed the threshold. Matches of rules in Log mode never count
Detections within minutes 20 in 10 min The threshold
First ban 15 min Each repeat offense doubles the duration …
Longest ban 7 days … up to this (at most 30 days)
Permanent ban from offense 0 (never) Ban permanently from the N-th offense
Also ban suspected scanners Off Also ban IPs that get many 4xx responses in a short time (path scanning)
Ban on bad behaviour immediately Off The gateway counts error responses (e.g. 404) per IP and site and bans as soon as the threshold is reached, without waiting for the next check
Exempt networks – IPs/CIDRs that are never banned automatically: office, monitoring, your own servers

An automatic ban covers an IPv4 address, or for IPv6 the client's whole /64 network (shown as for example 2001:db8:1:2::/64). An IPv6 device can switch between many addresses of its /64, so banning one address would not stop it. Detections from all addresses of a /64 count together. An exempt address or network inside a /64 keeps that whole /64 from being banned automatically.

Shared IPs

Where many people share one public IP (company or mobile carrier NAT), they can be banned together. Don't set thresholds too low, and add known shared networks to Exempt networks.

Manual bans

Ban IP bans an address for 15 minutes up to Until removed. In the list, Unban lifts a ban and Make permanent keeps it forever. From an alert about a repeated attack source you can ban the IP for 24 hours with one click.

Bans don't apply to the admin console (port 9443). To limit who can reach the console, use Network → Admin console allowed networks (see Users & sign-in).

Rate limits

A rate limit policy caps how fast one IP may send requests to a host. Create policies on the Rate limits page and attach one in the proxy host form (Protection → Rate limit policy).

Setting Meaning
Requests / s Average allowed rate per IP
Burst Extra requests allowed in a short spike
Concurrent connections per IP 0 = no limit
Status when limited Usually 429
Observe mode Nothing is blocked; would-be blocks are recorded in the traffic monitor
Exempt ranges IPs that are never limited

Tip

Start a new policy in Observe mode, watch the Traffic monitor for a few days, then turn enforcement on.