Insert manual CI in service mapping through pattern

[rimvydas_v]

I am trying to crease service map, however after load balancer we have TMG servers to which discovery does not have access (and probably will not have in nearest future). During mapping procedure connections from load balancer to TMGs are created, but these devices cannot be identified. Error message:

Failed to detect the operating system on host: x.x.x.x. Unable to discover host by IP address x.x.x.x. 

Quick discovery Shazzam returns closed ports.

What would be the best course of action add these devices into the map and continue with the rest of service mapping? 

[robertgeen · Accepted solution]

Unfortunately unless you are on Kingston release this isn't possible. You can create manual connections but you still have to have the ability to discover them (unless you are on Kingston release). In Kingston they added the ability to add manual CIs to service maps which is exactly what you need so unless you plan on going to Kingston this probably won't be possible.

I feel your pain as I've hit moments many times over where it would be of been so nice to just add a manual CI to the service map.

[rimvydas_v]

Thanks for the answer.We are planning to move to Kingston this year. For the time being I will try to use multiple entry points and jump to next devices in the flow.

Just for the curiosity how service mapping is different in Kington? I know that it will be possible to select $pre_cmdb_xxx valiables from cmdb tables. But the question is how initial identification will work to know from which ci record to take information?  

[Paul Ciarfella]

I don't see Kingston solving this problem completely unless service mapping and discovery support many more commercial devices out of box and can trace connections through devices that are not discoverable.

For example, we've got WAN optimizers and security appliances between entry points and back-end servers that ServiceNow does not support.   Even if we're able to setup proper credentials on these devices, likely SNMP only, we will still need to build and support custom patterns for these devices, and only if the device can return connection information via whatever interface we can get access to.

[robertgeen]

Yes I agree it can be difficult if there are many network devices and appliances between them to get everything as you have to have ARP cache and layer 2 topology discovery work perfectly. While you could drop them on their manually and build it out it negates some of the dynamic nature of the map which is exactly why people want to use it.

It's all about whether or not the information is there in the fashion that it expects. I'm not saying it would be impossible but it surely wouldn't be easy.