adam.beedell · SRE/Machine Learning Student · ✭✭✭✭✭

@"SteveK" Nicely done for actually producing the script! @"DavidVT" if you wanted to implement something this way, it'd be worth having a good think about the impact of disabled/enabled vs uninstalled services Just happened to have started poking around in SCM, and you get get alarms on software changes very easily, looks…

But yeah if you can hand a swis call to a dev that's definitely secure-er

Sans other solution, what I'd do would be to set a group with maximally strict limitations on it, to ONLY allow groups to be seen, an to disable all menu bars, and give a homepage. Setting a limitation down to one thing is easy with one product but increasingly hard as you get more products, there's a UI limit of 3, but…

Very odd. Really feels like there should be a "new" interface visible If you add the node again is it the same issue?

Is it a proper computer running the autoreloaded page? If so might be able to see what's memory leaking with f12 devtools Could also drop a monitor on the box to see if so. If it's a smart TV ir something less so. Got anything particularly custom on that page? If you can prove it's a memory leak, should he good for SW…

I think you're in the wrong product forum, as NCM and UDT arent part of SWO (thats the cloud-based-one) That said, if you give a minimum access group account then default homepage to a map and disable all other UI you'd get somewhere close to what you want, but if you want to be able to view statuses you can also view…

I dont know for sure that EOC's port updated, any luck on 17777?

Do you have current data for anything else on that device? (Say, the CPU chart?)

Oh interesting. You got a lot of these? 1) 2 alert actions with a wait 2) 2 alerts, one looking for the first being down, then one looking for the 2nd one being down (either slower, or AND looking for evidence of the first happening. EG, add an action to the first filling in a custom property, for the 2nd look for the…

~'WHERE switch status = up' perhaps

When it breaks, if you refresh the page manually, does it appear you've logged out/been timed out/etc?

There's a few ways you could solve this 1) Set the alert to reset after the this action is taken (or add an action to clear the alert), when it next triggers it'll go one down the list 2) Set a custom resetter action, pump the full list of broken services to that (this is a faff, you'd be doing a select ID where type like…

As written that applies to everything that uses that poller I believe, else a simmilar name attribute would work.

search "next poll in the past"

This made come off as "Yeah of course I checked that, duh!", but did you list resources, and find a new interface? (or anything)? If new interface, check that, someone changed the setup If list resources fails, agent has failed somehow, maybe reinstall or whatever If same resources show, interesting. (Make sure it actually…

If you updated to 17774 with 2024.2 then maybe that Ultimately if the website is still up then SWIS is working, if [You->Swis] isnt working, then the issue's somewhere between you and there. Think DNS, Firewall out, midpoint firewall, firewall in

You're using SQL, he's using SWQL, SWQL is better for this stuff long term because the SQL will break if SW change the database between patches Aside from that, the first 2 columns are easy, the last one can very based on vendors and equipment your environment uses (and sometimes what counts to the audience)

Tried what exactly? (again, cant read your screenshot)

Looks the same Should really use SWQL for this sort of thing If the results between the two queries are the same, but the last of servers doesnt match the CMDB list, that's pretty normal, and you'd need to compare the lists, add some stuff, maybe recategorize some other stuff, sometimes remove some stuff too.

Jokes aside probably the port's not open between where you are and where the MPE is

I dont know what your business hours are, this approach in the code seems unusual but could well be the best for you You could imagine most people getting by with a SELECT CASE WHEN WEEKDAY(GETDATE()) in ('weekdaynumbers') then '9-5' ELSE 'weekend' If you needed to incorporate stuff like local timezones for devices and…

You should not have to do this If you have to do this something is likely to be misunderstood If it's not misunderstood, something's probably wrong if something is wrong, you'd probably have to edit it in the DB, you can open a DB editing session from the main polling engine

This feels like an SCM kinda problem really. Events is probably the way to go, if no event windows troubleshooting makes sense as a path You could make a monitor in SAM that counts total services, then set that number to track difference. -x would be an uninstall, +x would be an install. Should work,

Which browser and autorefresher are you using? How do you authenticate the wallboard account? I've got some classics on screens using modernUI elements which seem to update fine at the moment but I've had a pain with em in the past, mightt be able to test some modern-only stuff. No promises I do find the settings above…

Try LIKE instead of =

I would expect them to not show as new once imported

The alert is looking for a gap, if you have a gap and don't want to alert, then you should increase the smoothing and/or disable that alert around downtime I do have grouped version of this alert, but the above is the more direct solution I quite like setting the alerting service to delayed start, this gets reset on every…

@"kpina" 's reply answers it better than I could @"surnaihai899" 

Mostly you can treat it as SQL, with a reduced list of functions, some auto-joins, and slightly different date functions. Any complicated join stuff can be learned as SQL first using those resources and then applied (though if you do this, as best practice cancel any unnecessary joins)

1. Whats the context from which you're referring to "onboarded devices"? I'd assume this meant Nodes, but maybe this is something someone said, I dunno. 2. Thing with an ID in the nodes table. I think this is maybe a licensing question? If so there's some cutouts for ICMP-only stuff, but then those are still nodes. 3.…