Skip to main content

Reliability

We take reliability seriously, we will describe how we set up our systems to make sure that you always get top-notch service!

NSToolBox IP information#

The IP information page (located at nstoolbox.cloud) is currently hosted on our cluster of servers, it is not dependent on a single data-center. IP information on the homepage (IPv4 & IPv6) and IP information on the IP pages are provided from three data-centers, your browser will request the information from two data-center at the same time, once you get a response, your browser will cancel the other request. Even if one of our servers (or a data-center) goes down, everything will still work.

Server Cluster#

The NSToolBox API & Dashboard are hosted on a cluster of servers, that will automatically fail over when required. The servers are not hosted in the same data-center, and not on the same network (ASN), and can thus survive a data-center wide outage.

DNS#

For DNS we use a mix of ClouDNS's anycast network & Hetzner DNS, we are not reliant on one DNS provider, so you should always be able to resolve our domain.

Uptime Monitoring#

Every uptime check will have at least three probes (server that will check your endpoint) assigned to it, every minute, every probe will fetch what monitors it needs to run, if it does not check in we will internally mark that probe as "dead" and a new one will be assigned to your uptime check until it is back up, so your monitoring checks are not affected by issues with our probes, nor will it cause false positives.

Email Monitoring#

Email monitoring is fully redundant, every mailer has at least 2 mailservers that will receive emails, if one mail server goes down, the other one will still receive the emails. Our checker will check both mailservers every minute, so your mail monitors are not affected by this. The mail servers are also hosted in different datacenters/countries.

DNS Failover#

DNS failover depends on our uptime monitoring system, it will run jobs on our cluster of servers, there is no single point of failure, if we lose a data-center, another one will take over.