Hello,
I am using the observium professional edition and I’ve been looking into configuring IP SLA on all my cisco devices to monitor our WAN. I have the following config on two devices to just test the waters
Destination device config ip sla responder
source device config ip sla 1 udp-jitter 10.34.36.2 16384 source-ip 10.34.36.1 frequency 300 ip sla schedule 1 start-time now
My question is what does the community do as a best practice for IP SAL configurations. As I have it configure now it’s send 10 packets every 5 minutes. This seems like a bad idea because what if there was a network blip between the 5 minute run time? If there was packet loss I would like observium to detect it, and send out an alert. Do I need to configure the ip sla service to constantly run? And if so how do I make sure all that data is there when observium polls it next.
—Jeff
Hi!
For the packet loss probes, only ICMP-Jitter operation supplies polling values to the SNMP MIBs, so if you want to measure packet loss, use ICMP-Jitter. The limitation is in the type of the probe for the ICMP-Jitter - not all servers in the Internet respond to this probe - for instance, 8.8.8.8 doesn't respond to it. For me, nice value for the packet count is 100, with 1000ms interval between them - you'll have info about packet loss for 100 seconds out of each 300-second polling interval, config for this looks like: (config-ip-sla)#icmp-jitter 192.168.1.1 num-packets 100 interval 1000 (config-ip-sla-icmpjitter)#frequency 300 Limitation of the ICMP-Jitter is that it is supported only on the routers, with classical IOS, none of the IOS-XE boxes seems to support it. Alerting config for this is at http://www.observium.org/wiki/Alerting_Metrics_and_Attributes#Cisco_SLA
For tcp-connect operations without `ip sla responder` (mentioned by you), use `control disable` config option: (config-ip-sla)#tcp-connect 192.168.1.1 80 control disable
Feel free to ask further if something'll be unclear.
participants (2)
-
Jeffrey d'Ambly -
Сережка Хомяков