banshee/network-collection/network/snmp/
system
The name and the time since boot of a switch, a router or a firewall by SNMP, to see a reboot or a device replaced.
par banshee1 installation2 versions
Name and uptime of a network device, by SNMP
Returns the name a switch, a router or a firewall gives itself, and its time since boot, in hours. Two watches are offered: a recent reboot, and a name other than the expected one.
It takes over the device's part of network/snmp/interfaces version 1, which mixed the device and its interfaces. Its outputs keep their names there, system-name and uptime; the uptime is now in hours, no longer in days, to see a reboot of the last half hour.
How it works
The collector reads two values of SNMPv2-MIB: sysName, the name of the device, and sysUpTime, the time since its SNMP agent started, in hundredths of a second. On a switch, a router or a firewall, the agent starts with the device: its uptime is the device's.
- A reboot shows as a time since boot shorter than the threshold of "Recent reboot".
- The name watch catches an address that now reaches another device (a replacement, an address reused).
The collector only reads: SNMP sends nothing but reads.
Prerequisites
- The SNMP agent of the device enabled, reachable from the poller (UDP 161), its access restricted to the poller.
- The view granted to the poller includes
system(.1.3.6.1.2.1.1), which every read-only view allows. - Read-only access, SNMPv2c or SNMPv3, set on the host or one of its folders (
protocol:snmp:*variables).
Outputs
| Name | Label | Unit | Per instance |
|---|---|---|---|
system-name | Device name | — | no |
uptime | Time since boot | h | no |
Watches
| Name | Title | Kind | Defaults | Checked by default |
|---|---|---|---|---|
rebooted | Recent reboot | below a threshold, on uptime | warning below 0.5 h | yes |
system-name | Unexpected device name | expected value, on system-name | no name given: it raises nothing until a name is set on the probe | no |
Limits
sysUpTimecounts hundredths of a second on 32 bits: it wraps every 497 days, which the watch takes for a reboot.- A restart of the SNMP agent alone, without a reboot of the device, also shows as a reboot: on a server, where the agent is a program among others, prefer the uptime collector of its system (
linux/snmp/uptime). - A reboot shorter than the probe's interval is seen only if the threshold is longer than that interval.