jbiggley ✭✭✭✭✭

Nice! Fortunately that query with the DATEDIFF only runs once per day as a stored procedure but I'm definitely going to look at other alerts that might be using DATEDIFF for sure.

I'd love to see some PS goodness.

That sounds like the most awesomest script! Did you do that in PowerShell?

Based on my experience, you shouldn't have to hold your breath a whole lot longer. If XenServer support is listed, I expect you'll see something either formalizing it or indicating it has been dropped in the not too distance future. My hard-earned (Canadian) dollars are on it will be formalized and announced in an upcoming…

Oh chrispaap​, you are such a tease. Global search...teehee! Now adding vSphere vSAN, Cloud monitoring (does that mean AWS & Azure??) and VMware vCenter events is absolutely drool-worthy. And I am very curious on how we can make the VMAN appliances go away -- or at least become less relevant.

Amen and amen. Or at least have an alert that would trigger an action to notify us of new VMs that met a certain criteria. (ie in a certain cluster, matching a certain name, etc.) 

I love the idea of grabbing the vCenter events. That means that the vCore teams can continue to manage with vCenter and their tools but we can align on the enterprise event management standards for getting those events into incidents. +1000 for Orion Global Search too

Indeed, but our script is a little shorter. We run this as an application component in SAM. The component is a PowerShell script running in Local Host mode using an account that has enterprise-wide AD rights. The component is set to go critical when the returned value is greater than or equal to 1. (Statistic 1 indicates…

Holy awesome-sauce! Way to go HolyGuacamole. I echo Leon Adato's analysis -- elegant and simple. Well done, sir, well done.

Indeed! I imagine you'll see this process shortly

Hope it helps!

Four years ago, I was asked to write a simple PowerShell script. It took me a week. The script was (Get-Process -Name *blah*).count I am living proof that if I can learn to PowerShell things, *anyone* can learn to PowerShell things

Trolling Master Class is now complete. Nice find zackm​!

I haven't tried but CPU queue length, if measured via a component as defined above, should be graphable. You can definitely see it in the component details for the application monitor.

Sorry, last few weeks have been chaotic for me. Glad you made some awesome progress though!

I haven't re-written the query but it should still work. The good folks responsible for the DB schema at SWI turned a bunch of the tables into views when they moved tables around. That should have kept everything working. That said, you *can* emulate this same alert logic without using a custom SQL or SWQL query now. Just…

Aha -- I figure it out. (I think) The instanceId field is the NodeID in either NodesData or Nodes. The ThresholdNameId is the foreign key to connect the ThresholdsNames table on the Id field. SELECT TOP 1000 [Id] ,[EntityType] ,[Name] ,[DisplayName] ,[DefaultThresholdOperator] ,[RecalculationNeeded] ,[ThresholdOrder]…

The thresholds are stored in the dbo.Thresholds table. I haven't tested the query, but it would look something like below although I'm a little puzzled as to why the Thresholds table sometimes has two or more rows for the same instanceID. Me-thinks there is some testing to do to figure out what is going on. Regardless,…

Are you interested in referencing the custom threshold or the dynamic threshold for CPU?

As a large customer, I can attest that the single installer definitely makes my life easier. I'm looking forward to unleashing that goodness on the NPM 12.1, SAM 6.4 and SRM 6.4 upgrades that were just released.

Ouch. Your upgrade path wins

So aside from stacking the installs (which makes me a feel a little dirty inside tbh) I do everything else in this list except for using the all-in-one installer on the APEs. (Awesome post BTW -- should be a pinned post IMO!) We do have a dev environment that replicates the installs in our prod environment (primary, APE,…

Well I tend to forecast a little longer for our change window so that we can get the business to support it. Extending change windows is nearly the unforgivable sin Our upgrade path is going to be NPM 12.0 > SRM 6.3 + HF2 > NPM 12.0.1 > SAM 6.3.3 > Orion HF as per the upgrade advisor. I'm hoping to gain some time by…

Yeah, monitoring network infrastructure is hard because of exactly the scenario you indicated. For that reason we only monitor "interesting" interfaces. Uplinks, server connected interfaces, etc. If you are an access point and want to know which switch port you're connected to -- good luck Bigger isn't always better. Or…

Don't be too jealous. I am staring down a 13+ hour change window in order to upgrade to NPM 12.0.1, SAM 6.3.3 and SRM 6.3

It's that whole NPM + SAM in a large environment too. Throw in 13K+ nodes plus all of the application monitors and suddenly you feel the need for a bunch of polling engines. Is 17 overkill? Probably

I swear I read 12K somewhere just recently. And stacking freaks me out. Rolling a new OS out was cheap for us so we opted for resiliency with 17 pollers on their own servers. Yes, lots of things to manage but an OS issue doesn't bork 20K elements

12K elements feels like it is too many -- although with beefy boxes like that you might be able to push it. We've targeted for 10K elements per polling engine.

Sometimes I re-read my comments and think "Wow, that is either the smartest thing I've said or the dumbest thing ever." Please, *please* make sure you design your monitoring strategy with security in mind. All of those ports? Inbound from *only* your polling engines for monitoring. Don't get lazy and use CIDR rules because…

Do you have security groups defined to let you do monitoring via ICMP and SNMP/WMI from your on-premise environment? WIth a Direct Connect you have the option to use those protocols as they aren't being NAT'd (except if you are using ELBs -- ugh -- see HA Service Monitoring in AWS via an ELB using PowerShell ) and leverage…