Autostatus is a network and server monitoring program. The project is designed to support large, arbitrarily complex networks of computers, and still provide feedback in a very timely fashion. To that extent, it has the following goals:
· Speed - Autostatus attempts to probe for network status as quickly as possible.
· Efficiency - Autostatus will not attempt to check for status on portions of the network it already knows are down, eliminating redundant checks.
· Accuracy - Autostatus presents as accurate a picture of the network status as it can from its perspective.
Autostatus has some features which make it fairly unique among network monitoring programs:
Autostatus can be given information about the path it takes to reach network segments, or other dependencies in network or server topology. Given this dependency information, autostatus will then query for status in the proper order in order to guarantee that:
· Nothing is queried before all of the items upon which it depends have already been queried.
· Nothing is queried if an item upon which it depends is down.
· Items are not incorrectly marked as down when they are simply unreachable due to a dependency failure.
This dependency resolution is, of course, fully recursive. The software will also warn about abnormal conditions in the configuration file such as cyclical dependencies.
Parallel status gathering
Autostatus uses the excellent fping program written by Roland J. Schemers III at Stanford University. (see the fping copyright below). By intelligently grouping services together for parallel testing, Autostatus minimizes the amount of time necessary to scan large numbers of networks or servers.
For testing TCP services, Autostatus uses a component called tcpcheck to perform parallel checks of TCP connections. tcpcheck was developed specifically for Autostatus, but has considerable utility outside of Autostatus.
Status Web Page
Autostatus automatically generates a status web page which reflects the current-known network and server status on a per-item basis.
Autostatus sends email when the state of a monitored item changes (based upon configurable thresholds). Using an email to pager program (such as hellpage, by the author, or various other paging programs), it is very easy to route notifications for each service to the correct party.
Multiple service monitoring
Autostatus can monitor both routers/hosts via ICMP messages, and specific services on machines via TCP connections. In keeping with the themes of autostatus, TCP connections may be avoided if the machine is not reachable via ICMP, or may be tried at all times.
Usage: tcpcheck < timeout > < host:port > [host:port]