mesverrum · Observability Architect · ✭✭✭✭✭

So you have a decent chunk of hardware on the system, have you been able to ask a DBA to look through and find what queries are getting deadlocked? Deadlocks are 100% of the time a problem. Something is wrong that you see consistent deadlocks, and since it doesn't seem immediately likely that it's a hardware deficiency you…

FYI, on the backend it seems like you can insert just about any SQL query into the definitions and it will execute as expected, but if you try to hit the edit in the UI you will find that it does not display them correctly if it is an operation that they don't include in the web ui.

I was able to get your original query working just by adding in a placeholder value to use in the group by/legend shows boxes SELECT DATEADD(dd, DATEDIFF(dd, 0, TimeStamp), 0) AS 'Time' ,COUNT(*) AS [triggered] ,'Count' as [Legend] FROM AlertHistoryView ahv with(nolock) WHERE EventTypeWord = 'triggered' AND (TimeStamp >…

IIRC you might be able to just throw your whole query in there as normal and just rig it up so that column 1 is an integer, maybe something like select 1, * from table In the past I remembered that it would basically just ignore anything outside row1 column1 (stat) column2 (message)

Do you have SEUM (WPM) in your environment? if not then just delete any lines referencing those tables

Yeah, I've done this stuff before but it is so few and far between for me I have forgotten the syntax. I've just been too lazy to log into my lab server and try to build a template using one of these to jog my memory, was hoping quick google fu would save me from work. I'll let you know if I come across the syntax when I…

Glad you figured it out, good luck

Do you have ncm? It there should be a built in physical entities report that I sometimes filter to just show the chassis model numbers. The tricky part is cisco is very inconsistent about how they differentiate that main serial across their product lines. I'll have to look it up when I get back to work but the script I use…

Try something like setting the select to select top 10 and see if that executes, if so thats uaually a good indicator the report is taking too long to generate (default time out is 5 minutes, but it will show an error on the screen if it exceeds that normally)

Ah nice to hear that got sorted out, when 4.4 first came out they seemed reluctant to say you would be allowed to run them together.

The agent is having cert errors because it is expecting to be able to use the cert that was used by the previous system. It needs a new cert. You could probably dig up some way to modify the files and replace the cert by hand but it seems infinitely more reasonable to just reinstall the agent from the new environment,…

Windows devices * Log on with an administrator account to the computer with the agent installed on it. * Open Orion Agent Settings in the Control Panel.If you get an error when opening Agent Settings in the Control Panel, navigate to the C:\ProgramData\SolarWinds\Agent directory, and Permanently Grant Access to that…

No, the ape installer will validate that the version being installed matches the version on your primary.

I just played with it, words like address or city are going to be ugly with a lot of false positives because there are a lot of matches for the term address, or words like capacity which are very common across solarwinds. Good luck Also it looks like, trying to hit multiple cp against multiple tables takes a brutal long…

I havent done that one with the sdk but i have done it with direct sql edits before, the groups are referenced in the database at containers, there are only a handful of tables referencing them. The SDK is the preferred way to make things like that happen if they have a verb in place for editing them. -Marc Netterfield…

Top x is after the where condition until it hits the cap, then it says it is done. But this thread on the same situation happening in plain sql seems to point toward your including a top statement could impact the execution plan in a way that makes it run more efficiently.…

I don't remember off hand if it gets purged, but if you aren't running inventory jobs then its likely that not all nodes have ever been inventoried. Maybe people have just been occasionally one off manually inventorying nodes one by one. That's where NCM is getting the data on those tables, no inventory then no interface…

Specifically with Linux/Unix based systems the failing is usually that lots of appliances and distros have a bare minimal snmpd configuration that doesn't give a lot of the good info people want. I have seen people customize the daemon to give them everything they want but it is not common. If you have SAM you can set up…

So parameter 0 would be the name of the poller your want to assign and param1 is an array of int32 netobjectID's, so for a node it would be a nodeid. A common problem people run into with the powershell module is if you pass an array with only a single value then it ends up changing it from an array to a the singular form.…

Did you apply some kind of filter to it, was it on a summary page or a node details page? Works as expected in my lab if I assign it to a specific node details page and tell it to use the current object as the source. If you are trying to pull it for all objects I wouldnt be surprised if it timed out since there are a lot…

Other possibility is that you need to change the method between get, get next, get table, and when you change those it sometimes requires you to tweak the OID and takes some poking around to determine exactly the correct way to structure an SNMP request for custom cpu/mem pollers sometimes. Got a screenshot of the results…

That feature has been out for a couple years now, Change the hardware threshold values - SolarWinds Worldwide, LLC. Help and Support As mentioned in other parts of this old thread, some versions of Cisco code were reporting bad thresholds through SNMP, so now I always have to verify when something has a nonsense threshold…

At a glance it looks like maybe your label is not set correctly for those pollers. I attached some screenshots from my environment for an example. You go to the UNDP tool in orion and right click the poller you need to work on and select Label then set the radio button to the "use labels from a table column" and set one of…

Unfortunately I don't have anything saved handy, I usually just spin them up on the fly with clients who have a use case for them. Things like linking from the node details page to all the IPAM subnets that it has addresses in via the nodeipaddresses to the ipnode table, then maybe correlate that to the dns entries from…

SELECT i.node.Caption, i.InterfaceName as [Interface], l.RemoteSystemName, l.RemotePortDescription as [Remote Port], l.RemoteIpAddress FROM Orion.NodeLldpEntry l, Orion.npm.Interfaces i WHERE l.nodeID = ${NodeID} and l.nodeID like i.NodeID and LocalPortNumber like i.Index

I'd suggest voting up these feature requests since they all come back to the same root issue. https://thwack.solarwinds.com/t5/LA-Feature-Requests/Log-Analyzer-Allow-searching-for-varbind-name-value-pair/idi-p/554301 https://thwack.solarwinds.com/t5/LA-Feature-Requests/Macros-for-Alerts-in-Log-Viewer-Analyser/idi-p/544060

Unreachable nodes are not polled at all until their parent comes back up.

There's not much we can do for you from here. You send snmp requests through and they send responses on port 161. Make sure your server can route to the IP, make sure the firewall is allowing traffic through, make sure you have the correct community string. Have you run a wireshark packet capture to confirm what is…

This is an example of the sql i am feeding into a custom chart, cleaned up to protect the innocent. select rad.ActiveUsers , rad.ObservationTimestamp , concat(n.CustomProperties.City, ' - ', n.Caption) as Site from orion.asa.RemoteAccessDetail rad join orion.nodes n on n.nodeid=rad.NodeID and n.caption like '%FWASA0%'…

You are kind of limited because you are relying on the built in function. That function requires the start and end times and doesn't make things easy for you in terms of changing the segmentation. I kind of suspect you will need to open up the function, steal the logic that they are using to actually do the 95th percentile…