mesverrum · Observability Architect · ✭✭✭✭✭

Netflow is a Cisco designed protocol that does not make any attempts to record latency, it's all about bandwidth utilization, so NTA also does not track latency. There is a feature in NPM called quality of experience that uses an agent and can display aggregated trends of DHCP request latency. Also in NPM there is a…

If an interface is not checked in List resources then you have effectively told Orion not to poll that interface for anything. You will need to add them to use UNDP pollers on them.

Depends on which specific models and brands of servers you have, but most of the newer IPMI systems report on power consumption. If you have the vendor software installed for the server type (Dell/OMSA, HP/Insight, etc) and you have enabled hardware health on the server you can check and see if the power consumption comes…

Not an easy way that I can think off from the top of my head, new-swisobject is pretty much like doing a sql insert, it only adds the thing you want to that single table. It never just figures things out for you, so you end up needing to set one up manually and identify all the tables that object starts showing up on, and…

To specify a single row in a table just add the row number to the end of the OID. This can be really tricky for OIDs that have a dynamic number of objects in their table, but if it is static you can just do that.

The company I work for has offices in Cork that service all of EMEA https://loop1.com/

So one thing to consider is that these logs were already being sent to your system in the legacy system, so the performance impact on your server isn't really that different, you used to get these events and they were being dumped on the syslogs/traps tables in the old db. They've just deprecated those syslog/trap tables…

If you have an APE then you should be able to get the info with a custom query resource or just loading the swql studio and running this SELECT EngineID, Name, Version, IsActive FROM Orion.ModuleGlobal -Marc Netterfield Loop1 Systems: SolarWinds Training and Professional Services * LinkedIN: Loop1 Systems * Facebook: Loop1…

Just to be certain, are your orion server and sql and all polling engines on PST as well? Orion doesn't like dealing with multiple zones

I find that creating nodes via the direct create actions is kind of hit or miss and requires a LOT of crawling the db to find where related changes might or might not happen. For my cases it has always been MUCH easier to script a discovery against my list of objects and then just let the discovery engine do all the leg…

So this is a sticky problem because normally orion doesn't update the list of available drives unless you run a discovery. If you do periodic discoveries of all your nodes you could build a report of all the drives that were discovered but are not in the volumes table.

What node you assign these components to is really irrelevant. Some people create a node pointed at the o365 gateway servers, but you could really use anything as a sort of placeholder. Looping in mrxinu , since he knows more than anyone about the setup of these scripts.

Depending on how eager I am to take advantage of features I tend to skip RC's. Yes they are supported and they will *generally* be stable but little quirks in them have occasionally broken features that I relied on. They usually get those issues hammered out with the first couple hot fixes but I like to let other people…

Not in the office to verify, but open up the ipnode table in swql studio and confirm that it allows updates, I remember when I last did custom stuff in ipam some of the tables were a little confusing on where you were and weren't allowed to make changes, as many of them are basically just views built from other tables.

I want to say that this resource is doing a join between nodes and volumes tables, in which case caption would be a column in both cases and you need to specify which caption you need, so I'd try volume.caption like 'c%', if that doesn't fly there is another field in that list that contains the drive letters by themselves,…

Can you right click the table details and see how large the table is? How many interfaces do you have and what is your polling interval for interface metrics set to? Do you know if the nightly maintenance job is completing successfully? It's not uncommon for that to be one of the largest tables in the database, and the…

I would just hold tight, the version of NTA in beta right now ditches the FSDB and moves it back into normal MS SQL, so all your efforts will be completely wasted when that hits general release in the next few months.

Not directly, you can call in variables for any column on the interfaces table in the database. One trick that you COULD do is to create a custom property, and write an alert rule that would update that custom property with the data for your custom metric. Or if you are comfortable in SQL directly you could write a trigger…

You are correct, the OOTB service monitor collects the cpu/mem/vmem/iops and there is no ability to control for it. You could roll own script based monitor doing something like a net service lookup, but it would only serve to make your life more complicated dealing with things like that. Best to just set all the thresholds…

Definitely a script kind of task.

Could be done with custom swql, would need case logic where if the object type is a undp monitor then the link would point to your summary page. I haven't had a need to make a query for what you are asking but it wouldn't be terribly difficult if you load up swql studio and look around to see what you need to do to make it…

I've done it in custom SQL reports. You essentially need to figure out the relationships such as this group of servers are all dependent on that UPS or this network interface. If everything goes down in a narrow window and we got no traps from the UPS then we assume the network is down, if the ups fires a trap indicating…

Since I don't have data like this in my environment I'm not entirely sure what the expected result is, but this seems like it would be correct to me, assuming that the run's are the same in both data sets, might have to play with the joins if thats not true. SELECT DISTINCT ncp.RUN, Count(*) AS PD , sbu.SBUPD FROM…

It would have to be SQL/SWQL because the web based writer doesn't give you a lot of options when it comes to adding values up, it only lets you do min/max/avg/count in most cases. I use a downtime report that looks at the event history and counts up the difference between the down event and an up event. I could imagine it…

You can use any monitored object as the parent of a dependency, so just add those tunnels to NPM and select them the same as you had been doing to routers. Just when you are picking the parent there will be a drop down that says node, change it to interface.

Every time a site is scheduled to be tested it uses one of these accounts, if all the accounts are already busy then the job basically cannot be scheduled until one gets free'd up. Having more accounts increases the number of concurrent tests your player can perform.

Do you have SAM? MUCH easier to just create a service monitor for your sql instance and then you can build a component alert that says if the sql process is below 80% cpu and node cpuload is >90 then trigger.

Check out the github for the SDK, it has examples of manipulating the groups. Particularly of interest I think would be the create group with parent verb. Back when I didn't even know powershell I was able to take those examples and build a script that walks through a list of custom properties and nests groups inside each…

Last time I had to do this I built a script that just looped through all the agent based nodes and redeployed the agent with my new engineid, i was using the DeployToNode verb, but it also looks like the AssignToEngine verb could work too. Orion.AgentManagement.Agent | Orion SDK Schemas

Nodes being polled via basic polling will not be storing disk space metrics, those are features of vman so if you do build the custom report it will be missing the data from the unlicensed systems. If you happen to have an unlimited license in SAM you could try to pull in the vcenter performance counters from there, but…