This discussion has been locked. The information referenced herein may be inaccurate due to age, software updates, or external references.
You can no longer post new replies to this discussion. If you have a similar question you can start a new discussion in this forum.

Does Solarwinds allow to check the ICMP status of the interfaces hosting IP address

Hi All,

Have been looking about this for quite sometime but didnt find any thread related to it so posting this.

Solarwinds polls the network devices via SNMP to get the performance related details and ICMP would be for getting a up/down status. Now many interfaces also will be having an IP address attached to it.

So is there anyway to check the status of it and alert? I checked with the existing options but there is no option for IP address in dropdown 'alert on object'. Neither under interface its showing anything details for IP address.

I have seen this functionality in NNMi tool wherein by default it provides this for the device added in monitoring. U can view the status of each IP address and also alert on it via ICMP protocol.

Trying to see if that can be made available in Solarwinds as many of the customers would definitely need this way of alerting apart from just SNMP admin and operational status for an interface.

  • Just want to make sure I understand what you are going for... you want to get the status of your interface via ICMP?

    If you are monitoring via SNMP you can get the status of the interface via SNMP.

    I'm not really sure your question is clear. Can you clarify a bit more?

  • I am referring to interfaces that host IP address. So I would like to alert

    on these IP addresses. This would happen via ICMP and not SNMP as we trying

    to check if that IP is responding or not. SNMP will alert on the interface

    admin and operation staus.

  • In order to do that, you would have to add ICMP only nodes. For instance,

    -RouterA has 2 interfaces

    • fa0/0 = 1.1.1.1
    • fa0/1 = 2.2.2.2

    add routerA as SNMP node to get the overall status and SNMP details of the device (loopback interface preferably)

    then add routerA_fa0-0 as ICMP to track that interface

    then add routerA_fa0-1 as ICMP to track that interface.

    just keep in mind that the ICMP will be from the perspective of the Poller, so if there is a routing issue that allows the poller to reach that particular interface and not your users, you will get false negatives and vice versa

    you could use IPSLAs and monitor via VNQM.

    I can think of a few use case would the interface be up but not reachable via ICMP, however, I monitor those differently. If you give some examples of what you are trying to monitor for, that might help.

  • Adding it as different for each IP takes lot of work. Plus it has to be

    resolvable if m not wrong. What if that's not possible? Plus your licensing

    goes double. For instace If I have SL500 and i have every node where I have

    3 interfaces with IP address attached then I have to double up my count and

    go for a higher tier for license.

    Yes I have used IP sla method, though not in solarwinds but in other tools.

    And I am aware that it allows to alert on probes which could be connected

    to users or applications.

    For few of our customers here who use NNMi, they request us to alert on IP

    address and interface down separately. They are interested to know when

    icmp is down on those IP address. And since I am evaluating solarwinds so I

    am trying to figure out if it's really required!!

    Maybe we can request some network experts here to give their ideas and

    experience in terms of best practice for monitoring. Coz every tool will

    have it's own design for setting up monitoring but on the network level we

    would not know.

  • EDIT: So after looking it up, I am apparently wrong.  Nodes do affect your license count.  I don't think it always used to be that way because I swear that it only used to mention Interfaces as the only licensing element, but maybe I'm wrong about that too...

    ICMP-only nodes don't count against your licensing in NPM.  (Fun fact, WMI Interfaces don't count against your NPM License either if you have SAM.  Only SNMP monitored interfaces count against your NPM license.  Agent monitored interfaces probably do too, but I'm not sure about that.).  Anyways, if he monitors a switch/router via SNMP that switch router has 20 interfaces he monitors and then 5 of those interfaces have IP's that he also wants to ping, bringing those 5 IP's in as ICMP-only nodes won't increase his license count at all.  He could monitor 5,000 ICMP-only nodes and they wouldn't count against his license at all...  It's just a lot of work and administrative overhead to keep up with and I wouldn't recommend it unless absolutely necessary.  It would increase the element count for the polling engine and bring the polling rate up a bit, but honestly ping-only nodes don't increase a polling rate by much, so you could monitor a large amount of them and still have plenty of headroom there.

    Orion does tell you if a SNMP monitored interface goes down and you can include the IP address of that interface in the alert message if you want.  Now, obviously a SNMP interface can be up but for whatever reason the IP address assigned to that interface can't be reached for certain reasons, so I guess there is a valid reason for monitoring that IP via ping, but I would only do it if it is of vital importance.

  • Correction here: ICMP-only nodes definitely count against your node licensing in NPM. If you use ICMP though you won't be able to monitor any interfaces on that node so it won't draw down on your NPM interface count. The node and volume licensing is cumulative from your installed Orion modules (e.g. NPM SL500 and SAM AL150 installed means you can monitor 650 nodes, 650 volumes, 500 interfaces, and 150 component monitors).  If you have any unlimited licensing, you have an "unlimited" number of nodes and volumes you can monitor. So if you have an NPM SL500 and a SAM ALX, you can monitor an unlimited number of nodes, volumes, and component monitors but still only 500 interfaces with SNMP.

    pastedImage_0.png

    As to the original issue, I agree with the previous suggestions - either add the devices in as ICMP only nodes or create IP SLA operations (assuming Cisco devices) and monitor the status of the operations.

    NPM gets the up/down status of the interface from the node itself with SNMP. The device knows whether the interface is up/down. You need another system (e.g. SolarWinds or another device running IP SLA) to check the status of the IP address because if you ask the node to check it's own IP, most of the time it will be up because it is checking reachability to itself.

  • Yeah even i was shocked to see the comment stating ICMP only nodes will not consume license. It has to becoz the structure is built in that way. Elements i can understand as it will work based on what we add to monitoring.

    Regarding adding as ICMP only its overhead task plus DNS resolvable requirement is another problem. IP SLA is more suitable for CISCO environment but what about other vendors?

    So my intention is since Solarwinds is already detecting the IP addresses; if they can enable the option under the "Alert on" drop-down for IP address then that would solve the problem. This wont require change in terms on how we are polling the device or IP i believe.

  • If you're just looking to know if those interfaces are up I would make sure you add tunnel monitoring to the interfaces (i.e. BGP status).  That would tell you if the interface is reachable and not add any additional licenses.

  • while this can be taken as an option but it cannot be used for all types of devices....in some cases the environment doesnt have possibility to configure BGP at device end due to their network design.

  • If one of these IP addresses on an interface goes down, is anything logged in syslog? If yes, you can alert using the syslog message. If not, then go with the above suggestions emoticons_happy.png