mesverrum · Observability Architect · ✭✭✭✭✭

Almost sounds like you just want a pretty version of the alerthistory table?

Everything you describe is actually very doable. For the alert action I would have it launch a powershell script that restarts the services, then maybe set up a 5 min escalation where if the situation has not improved then it emails you if the problem has not improved. I also have created several variations on load…

You probably wouldn't time out if you just ditched the order by. I'm fairly certain that it's not an indexed column in the db so it is slow sort or filter against it. You may be able to get the chunks by using the "with rows x to y" syntax

I wrote this a couple years ago to deal with a client issue, and a lot of the code would apply to your request. Some of it doesn't, but you can pick through it and see what you need. Note that it uses direct SQL edits for some changes because at the time those tables werent available to edit in SWIS, that may have changed…

You don't have to keep the config archives locally on the pollers. Just set them to all write to a UNC path for a file server that already has redundancy and replication set up. Solves the problem of them being scattered, and piggy backs off the fact that DR file replication should be a solved problem already so you don't…

So from a normal computer how would you check if your mobile app works? Is there a url that could be tested, are there servers that can be logged, is there some kind of logging happening somewhere?

In my alert messages I use this syntax to make sure the messages notify me about any dependency relationships that could be an issue. It lists of the name of each dependency relationship that includes the node as either a parent or child. Parent of the following dependencies: ${SQL: select isnull((SELECT cast(Name + ' - '…

An important distinction to make is that the query from @ciscoengineer will be showing interfaces with errors in the current hour, where your method shows interfaces with errors in the previous hour, because he is counting them from the error historical table basically as they happen, the "this hour" fields on the…

Yes that is all accurate. APE HA clusters work exactly the same as MPE HA clusters. You did leave off the actual licenses for your MPE modules such as NPM/SAM/etc though.

If you want to chart it out over time then you probably need to be looking at this table instead Orion.Packages.Wireless.HistoricalClients There is not table actually explicitly tracking the number of connected clients, but that table does show the start and end time of client sessions. Yours ago I wrote this query to get…

The charts always need 3 elements, one must be a number to be charted out, then we need a time series value, and lastly a label of some sort for the group by/legend. None of the fields you listed are actually a number. N.CustomPollerStatistics.Status is typed in the DB as a string, despite the fact that maybe for your…

Yes, a shell script using the linux script monitor is likely your best bet, did you check if there is already one to do what you need on the content exchange?

Good to have you back in the mix, you gotta get into the MVP slack channel more, that's where the action is.

the name tells you, 2020.2 would be scheduled to be released in the 2nd quarter of 2020.

I believe the group details view uses ${id} instead

not from the GUI options but it could be done with sql/swql

Any chance that the ASA poller stopped working on that node for some reason and it's showing stale data? I recently had a vaguely similar situation because someone had forgot to update the ASA service account on some of our nodes after the password got rotated.

No you only have to build it once, the SWQL example would create a link for every distinct city in my example, but you would use whatever properties you have in your environment. Does the client view have any kind of account or view limitations on it? Those would work automatically to cut this widget down to only display…

Thing I like about perfstack is you don't have to build one for every time slice, you just build one and then if people want a different slice they can just click the time selection at the top of the screen. The example in this post is basically exactly what I did. Build out a single example with the charts you want and…

not wluther, but I can say I've done queries similar to what you are describing before to build out lots of perstack views quickly. Tagged a property on a bunch of routers, then tagged the primary and secondary network interfaces appropriately, left joined each category of interfaces to the nodes, turned the whole mess…

I've always found this method from @"AlexSoul" to be really powerful in terms of being able to do the dependencies via automation. His explanation is pretty good and the method is easy enough to follow. If you are into scripting you can build up a lot of automation from there. I tweaked that model for my use cases so I…

Absolutely hate this bug, bit me several times now. 

Last I saw the DPA api only really supports a narrow set of commands, mostly relating to admin stuff like adding new instances and deleting old ones. Traditionally I have seen people just going directly to the database of dpa to do bulk work. Considering that many DPA user are DBA's and that is their natural playground.

1 ) nothing that I ever saw, the baselines are calculated as part of the db maintenance job so maybe if your maintenance log shows errors like it timing out that could be a thing. New app monitors don't get baselines until they have been in the system over a week also. Maybe if they removed a template and readded it. 2)…

Maybe it doesn't like it because testing api is not a valid status? This node is set to external/ or unmanaged status I assume because anything else will cause it to poll and override whatever you are cooking up.

that is the volumeid for the drive you selected

Most people I've known just set their service account to not expire. If they have an infrequent password rotation policy like once a year they just schedule it and plan on manually updating it, if they have a more frequent rotation then they usually leverage the api since most passwords in Orion can be changed there…

I'm going to discourage you from building this as an alert, just because every place I have ever seen someone set this kind of info up in an alert they get annoyed by all the short term blips and spam it generates. You are much better off using a report to spot check these periodically, on my admin dashboard I have a…

Thanks for pointing this trick out out, I co-opted it and started using this method to replace all the links in my existing collection of custom SWQL as soon as I saw it. Figured I should post back an example so people can see an additional use case for this idea. So I think the default cpu/mem and latency/loss gauges take…

There is no mechanism for that built in. Whenever I have worked with large sets of Atlas diagrams I've just had to rely on map naming conventions to keep things organized. Probably also worth keeping in mind that atlas diagrams are on the way out in favor of the newer Orion web based maps, so if youhave a big collection…