marcrobinson · Observability Detective · ✭✭✭✭✭

Agreed. You need something to at least ping the primary server until it is online. I would recommend checking it after patching anyways. Windows patches 'never' cause problems. 

I just ran that upgrade, also switched from legacy licensing to HCO. Licensing was the kicker, otherwise the upgrade went fairly smoothly. I have multiple polling engines and web servers. Otherwise, take a look at some of the other thwack postings. Based on general responses in thwack - the 2023.1 has been better overall…

I think that the DB Maintenance error reporting bug from 2022.4 remains. There is one known bug where container entity maintenance misreported. You may have stumbled across one of those errors. Please post if this is the case or if it is a new one. 

Actually, I would argue that 2020.2.X was stable and the 2023.1 GA appears to be stable. I just ran the upgrade from 2020.2 to 2023.1 and other than some weird licensing and windows permission errors during install, it went smooth. I am still validating overnight jobs with NCM. Looks like my mail settings needed to be…

Ran into similar issue with unable to access folder to delete files. I checked and the folders were already empty. Turns out I had to reboot and rerun the installer. Since the uninstall already completed, I was concerned, but the installer picked up right where it needed to be and complete the install. I did call support…

Interesting. Have you looked at using the API to modify/set a custom property, that then triggers an alert? Its roundabout, but would allow you to have the alert actions reset that property too. Its also pretty flexible as all you have to do is set the properties you want and now you can effectively trigger them…

There are some other benefits here such as simply separating the user/scripted rest api traffic. There have been a few threads about controlling access to the api, and this can definitely be used to observe where the interactions are coming from and limit what machines (yes the discussion has gone that far) can access that…

1663027 – net-snmpd double free or corruption error (redhat.com) Here is one of them. 

You would have to download all checks and tag information, then create a script to run api calls against a given group of tags/check numbers from your downloaded list... Ugly to say the least. If you do create it, and are willing to share that, I am sure that the gods of thwack would smile upon you. 

I do similar to track what is going on. If I see a lot of muting - I usually expect some systems to go down for maintenance. Audit logs can also effectively track node adds and node removals. But locking down a user to restrict the API and powershell is limited at best currently. I wish more auditing was done on the api…

The tags show up in the Check Detail for uptime checks. https://api.pingdom.com/api/3.1/checks/<checkNumber> BUt that would be one request per check to get that information. So using tags as a filter? meh not likely from what I have seen. 

What version of the net-snmp client are you running? There are some known issue with RHEL8 and some of versions. Mostly they seem to cause snmp to crash. 

Please consider bumping this request too. Basic Auth is not enough. SolarWinds API should support OAUTH2 and Bearer Token, and not just basic auth. - Feature Requests - The Orion Platform - THWACK

I have noticed this behavior with restful calls on the api too. They have to point at the primary poller. Its not a new limitation based on my experience. 

Educating the users to change the polling engine if they are manually adding them is probably best. I am not aware of a method to force users to pick a different one. I suspect a an alert with a scripted action to move a node a different poller based on a given custom property and polling engine id would work too.

Swugtember just does not have the same ring. It is probably for the best as I am sure I would not survive either. 

Agreed @JaroslawLadyga that windows firewall settings should be looked at. Consider using wireshark to monitor the traffic from the new polling engine - if the wmi call is refused by the target server, then its definitely firewall/security. There are some collector logs and DCOM errors in windows event logs that show up…

I think I smell a bigger thread and a feature request. I have been concerned about systems that automatically spin up servers and shut them down too. Makes discovery very interesting. At least containers are handled by the container monitoring - besides an ouchy involving perfstack non support (not sure if that is fixed in…

Custom Properties and even Alert Schedules. You can mix properties that you create like - AWS_AlwaysOn: Yes/No - that could be used to determine if a cloud based server should always be available/on and have the alert scope look at that property. In other cases, alerts may ONLY be desirable during a certain timeframe or…

Hmm looks like orion.nodes entity has some of this. Look at the IOSVersion field, then do a similar Case function from the example above. SELECT TOP 1000 NodeID, Caption, NodeDescription, Description, Vendor, IOSVersion, MachineType FROM Orion.Nodes Where machinetype = 'windows 2019 server'

Hmm, this report might not work as I believe that the assetinventory entities have the information, which I think are from SAM, not NPM. But here it is: SELECT ais.NodeID , n.Caption , ais.OSVersion , (CASE ais.OSVersion WHEN '10.0 (Build 20348 Multiprocessor Free)' THEN 'Windows Server 2022 ' WHEN '10.0 (Build 22621…

SAM is not a bad choice actually. It will work until you bring the system down for maintenance. Otherwise - you are correct that an external monitor would be preferable. If you have HCO - spin a single node license instance and have it monitor your production HCO instance main poller/web portal. 

@"dunky2k" Great Tip, that is a simple solution to the problem. Its one I had not thought about, thanks for sharing. Otherwise, I agree that you cannot add a modern dashboard directly to a tab. They modern and classic dashboards are separate, which is killer as I REALLY want to put the KPI widgets natively on classic…

Agreed. I ran into the same issue and renamed the files - reran install and no issues. 

Pure has an API that you may want to leverage also. SRM also leverages the API. There is also SMI-S - but that is more SRM specific. Introducing the Pure1® REST API | Pure Storage Blog

Yeah I dont have a better answer for you. All I can hope, and based on your comment above is that the offline installer will work better. Wondering if there is something with the online installer and 2023.1 downloading slowly. 

I hate to say this but try pausing your anti-virus during the installation. It took hours on a dev box for me. I was using the online installer. 

Came up with something similar, but WISH I HAD READ THIS FIRST. Thank you for posting as this is cleaner than my swql. 

You would need to grab the Offline installer for the 2022.4 GA version. I ran a dev install and noticed it was 2023 also. AND I had some bizarre issues. Ugh. 

Ok, I need to stop working. I am seeing things. The files are there, but the names are off. Here is what I noticed upon closer inspection. Observability-2023.1.0.129-Observability-2023.1.0.129-OrionInstaller1.cab Update: Renamed files (there are 21), clicked on retry, and it completed. FYI: I used the online installer if…