Monitoring Servers¶
In order to properly function, our monitoring servers need access to your services that should be monitored. With ICMP and HTTP checks, usually there's nothing special to consider, when monitoring publicly accessible services.
However, if you're running a protected environment that is built upon IP allowlisting, you need to explicitly allow access from our monitoring servers, because otherwise Argus will show your services as being unavailable, although they are in fact up and running ("false positive").
IP Addresses¶
Our monitoring network is dynamic and constantly evolving. We therefore strongly recommend automating the process of allowing access to your infrastructure from our monitoring servers.
For this reason we're exposing a publicly available API endpoint you can use to fetch a list of all IP addresses (IPv4 and IPv6).
Sending an HTTP request to that endpoint will return a text/plain HTTP response containing a list of CIDR networks to add to your allowlist, separated with a default line break. While today, we're only returning /32s for IPv4 and /128s for IPv6, this might change in the future and your implementation should respect the CIDR notation and not just strip the subnet's size away as this might lead to issues down the line.
Dual Stack¶
All monitoring servers are connected to the internet with IPv4 and IPv6 addresses. When a check is being executed against a hostname, that hostname is getting resolved each time the check is being executed. If a hostname resolves to both, IPv4 and IPv6, IPv6 is being preferred over IPv4 as it's the more modern standard.