mesverrum · Observability Architect · ✭✭✭✭✭

Yes, there are several ways to do it. One widget that can get you there is called the custom object resource and you could add one for for each interface or do more than one at a time if you like, or you could use the custom chart or custom table widgets, another option would be to save each interface that you want to…

Instead of using the out of the box top xx dashboard widgets you could just use a custom table widget, or roll your own SWQL custom query that only displays nodes that are violating their thresholds, and then just adjust the thresholds of individual nodes or disks instead of having a static 80% threshold across the board.…

No, "is equal to" is functionally a sql = operator which does not allow wildcards. I think you could do "contains" and achieve the results you are looking for there, or do something like starts with ch- and ends with -n

groupname isn't actually a column in the db so you can't WHERE against it, the correct place and syntax should be to push line 37 down and before the ORDER BY include something like WHERE c.Name like '%APS:%'

Try changing line 20 and line 27 so that instead of addmonth( -1 etc so they say addmonth( -12 Worth mentioning that doing reports that go back a year very often causes time outs, and if that's the case there isn't a whole lot you can do short of getting into the weeds with the database directly instead of trying to…

I like this, opens up some cool possibilities for dynamic dashboard generation as long as we can always tie the widgets we display to a common entity type.

I don't have an itemized list handy, because it's very long, but after an upgrade in my own environments the simple answer is that I check every screen in the GUI that we use, test every feature we use, and validate every script still runs. Especially the last year or two that last part has been a real problem for me…

The alternative approach would be to go diving through your IIS logs, but that is really painful unless you have a good centralized logging platform to be able to search through a ton of logs across all web servers at once.

It's kind of a chore, but if you don't use the precompiled version of the orion website you can modify the footer of the page to display the hostname of the webserver a user is connected to. The main nuisance is that every patch and upgrade SW tries to sneak you back to the pre compiled version and take away your hacks.…

SWQL doesn't have a cast function to force the data type, but you can do some tricks to make it happen such as: Select 0+myField from someTable ^ Doing mathy things before the column name gives SWQL the expectation that we are working with numbers. You'll need to make sure your where condition does a good job of filtering…

What do you mean? The RTO/RPO targets are always "Now/absolutely zero data loss" aren't they? Money is no object... until it is.

Haven't tested it myself, but your testing sounds believable and lines up with other quirky behavior I have seen over the years. I think one easy fix would be to just make it so one of your conditions is that the server status is not down, but like you said that doesn't change the fact that the time of day stuff behaves…

Yep thats exactly the kind of query i had in mind, looks good.

The size of the database has a few factors, some of which you can control pretty easily and others are more complicated to manage. * Number of objects being monitored - You probably need to monitor all your nodes, but when it comes to interfaces and volumes there is usually room to do some trimming here and remove anything…

I don't think it makes a big difference once way or another as long as you end up putting whatever credentials into your secret management tools so future admins can get to the passwords as necessary. It doesn't really provide a meaningfully different risk mitigation, but if it helps you sleep at night you can use 3…

If you figure out the details of the json formats that orion maps use to place things yes this is within the realm of possible things, but I've talked to a lot of people about map automations and so far nobody has built it. But there is also a question around what is your python going to be adding that Orion isn't already…

Ran into that recently when working with a new version, the format you have used to work but nothing I tried was able to trick the SQL variable thing into accepting it in the version I was working with. Will love to see if someone comes up with a solution that works in the latest releases because I smashed my head against…

I don't believe there is a native GUI based way to trigger a change alert for UNDP values, there are a couple places in the platform where they do have that available OOTB but I dont think UNDP is one of those. What I would probably do is cook up some custom sql/swql to just count the number of unique values for that value…

Most SWQL queries related to UNDP's are going to start out at Orion.NPM.CustomPollerAssignmentOnNode You can find a lot of example queries to use as a learning reference if you search for that table name in Thwack thwack.solarwinds.com/search

That's one of the strong arguments in favor of buying the Additional Web Server module, it's dirt cheap and allows you to offload the web stuff to a different server than the core app functionalities. I can't tell you how many times over the years I did something aggressive to one of my web servers in a LB pool and just…

Sounds about right to me, try installing NTM on a different windows box.

At a glance I don't like how your CMS join only associates nodeid to cms entityid. There can be many entity types that are not nodes and that could explain why you are getting unexpected results. I'd suggest adding more conditions to line 7, IIRC it should be something like AND CMS.EntityType = 'Nodes'

So from my perspective you are kind of falling into a trap by thinking you should use SAM to query the data and then cobble that together with an alert or some other trickery. I say that because if you have the relevant scripting skill to do that in SAM then you would have a much easier time just running the script as a…

Makes perfect sense, its hard to find people who want to do the actual work of running a boat so for now I haven't even bothered to buy my own boat. I just put it out there to local sailors that if they want to take it out they can call me and they keep my available free time pretty well covered. Probably in a couple years…

There are events in the audit log for users who signed in and certain activities in the pages. Not 100% of everything gets logged but enough to give you a ball park sense of how many people are logged in. If you want something more specific you can check the windows IIS logs, they are quite a bit more verbose, although…

Sailing But when I'm not trying to be on vacation I've been looking at new sharding and scalability technologies, saw some pretty interesting things at Reinvent this year that caught my interest.

to do anything more specific youll probably end up wanting to do a custom swql widget with some where conditions, the ootb table widget is kind of limited. You can go in there and uncheck each row that you dont want to display but i cant guarantee that the row id would always be the same across all switches for what you…

When you are creating a UNDP there is a section where it asks you about the format of the OID being polled and you can create enumerations, i don't have a screenshot handy but if you edit an existing undp I remember its about in the middle of the screen. The enumeration screen is basically like a spreadsheet where you put…

You are going down a very complicated road with this angle. One of the things to keep an eye on when it comes to alerts is to be aware of the type of entity the alert targets. Right now you are describing having an alert on a Node, but wanting to display variables relating to its child interface. But you already hinted at…

Seen this many times. Is there a reason that you are querying the esxi hosts directly instead of polling your vcenter? It doesn't solve the problem exactly, mostly it moves the issue but there is a lot less risk of having a problem if the vcenter CPU consumption is high compared to having it happen on each esxi host. 1…