ā04-23-2014 01:54 PM - edited ā03-04-2019 10:50 PM
Hey -
I would like to create as IP SLA for our remote Internet circuits that will trigger an alert when not working. This way we will be alerted even if the Outside interface remains up and pingable. The architecture is such that if the Internet is down at a remote site, all Internet traffic from that site will traverse the WAN pipe.So, we might never find out that the Internet is down at the remote site.
Should I create an IP SLA that is constantly pinging an address and trigger an alert when that IP SLA fails?
If so, would it be bettter to ping a known public IP address (4.2.2.2) or one of our own and configure the IP SLA operation only to ping the address via the Internet circuit?
Or, is there another way I am not thinking about?
Thanks, Pat
ā04-23-2014 10:19 PM
Hi Pat,
Understand you would want to monitor the internet link of the remote site. If you are using BGP with the ISP then BGP trap is a better option else would suggest IPSLA to one of your own IP.
Regards,
Sathvik K V
ā04-23-2014 11:51 PM
Hello Sathvik,
The bgp neighborship could be up - and internet still down, I'm not sure but enabling bgp trap may not work in this case. I think we want to achieve full internet connectivity monitoring from the sites not just bgp specific events.
ā04-24-2014 01:31 AM
Bilal,
True, if he wants end to end connectivity then IPSLA is a better option. I thought he wants to monitor only the remote site internet connectivity.
Regards,
Sathvik K V
ā04-24-2014 12:35 AM
Hello Pat, this is a good idea, however perhaps one small problem which you may have thought about already.
Lets say we have site A with internet circuits, bgp peering with ISP etc... If you have a backup route for all your internet traffic through your WAN, in theory the ping will never stop working even when the internet links are broken along the path.
So I would possibly do this...
One problem with the static route though pointing towards your bgp peer, if the interface was to go down, I believe that the static route will no longer be in the routing table, so this is a down side. But then one would assume you are monitoring the interfaces themselves. Therefor you have the option of pointing the static route at the interface for exit point.
R1(config)#ip route 8.8.8.8 255.255.255.255 x.x.x.x name ##IPSLA-VIA-BGP##
R1(config)#ip sla 1
R1(config-ip-sla)# icmp-echo 8.8.8.8
R1(config-ip-sla-echo)# frequency 5
R1(config-ip-sla-echo)#exit
R1(config)#ip sla schedule 1 life forever start-time now
R1(config)#
R1(config)#track 1 ip sla 1 reachability
R1(config)#ip sla logging traps
R1(config)#ip sla reaction-configuration 1 react timeout action-type trapAndTrigger
R1(config)#ip sla logging traps
R1(config)#snmp-server host y.y.y.y version 2c abc
R1(config)#snmp-server enable traps ipsla
I haven't tested the above configuration but I'm fairly sure it will assist you in achieving a complete solution
hth.
ā01-03-2018 06:52 AM
You can also use 'bool' in your ip sla and use several IP's to monitor with and/or. Check this link out: http://packetpushers.net/using-ip-sla-delay-feature-to-safely-monitor-lossy-links/
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide