Showing results for 
Search instead for 
Did you mean: 
Create Post

THWACKcamp 2017 - Orion at Scale: Best Practices for the Big League

Level 18

Anyone who is having issues with performance or considering expanding their deployment has had to wrestle with the question of how, exactly, to get the performance they need. This session will focus on maximizing performance, whether tuning equipment to optimize capabilities, tuning polling intervals to capture the data you need, or adding additional pollers for load balancing and better network visibility.

In the "Orion at Scale: Best Practices for the Big League" session, Kevin Sparenberg, product manager, SolarWinds, and Head Geek Patrick Hubbard will teach you best practices for scaling your monitoring environment and ways to confidently plan monitoring expansion. They will focus on maximizing the performance of your expanded deployment, and more!

THWACKcamp 2017 is a two-day, live, virtual learning event with eighteen sessions split into two tracks. This year, THWACKcamp has expanded to include topics from the breadth of the SolarWinds portfolio: there will be deep-dive presentations, thought leadership discussions, and panels that cover more than best practices, insider tips, and recommendations from the community about the SolarWinds Orion suite. This year we also introduce SolarWinds Monitoring Cloud product how-tos for cloud-native developers, as well as a peek into managed service providers’ approaches to assuring reliable service delivery to their subscribers.

Check out our promo video and register now for THWACKcamp 2017! And don't forget to catch our session!

Level 20

I'll be there for this!  This is exactly the kind of thing I'm always dealing with.

Yes, I am also very interested in this topic, see you there.

It's nice to see Thwack facilitating SW deployments for the bigger customers, or those who are growing, or who clearly see an organization's future is more efficient with better and more-complete monitoring.

Level 13

"Take Me Too Your Leader"


I'll be there, but I suspect that this information will be valuable for builds at all levels. IT always grows and seldom, if ever, shrinks - so building with growth in mind is always a best practice.



Level 21

I have found that Orion actually scales pretty well.  My biggest problem has been that I have to pay for additional polling engines versus a per-node model. I understand something like this is available now but I haven't had a chance to look into if it makes sense considering how invested we already are in the model that we have.

I discovered the hard way that EOC isn't a good fit for my environment, and that we can scale our pollers pretty well across 300 miles while having them report to just one instance of NPM--which is ideal, in my opinion.

Where scaling isn't ideal is when a single chassis switch can consume 1300 elements by itself.  I'd much prefer that SW (somehow) enable me to monitor all my nodes and their ports without having to grow to a second instance of NPM.

That's what EOC is designed to handle--multiple NPM's aggregating many pollers.  But EOC isn't nearly as fast or responsive as viewing the original info on NPM.  And having to buy multiple NPM's and view them all separately isn't affordable or effective.

I guess I just need a smaller organization . . .

Level 21

I haven't tried EOC but I have heard similiar concerns from other folks that use it.  I am fortunate enough to have not scaled beyond a single Orion deployment.  My scalability has mostly just been geographical (including public cloud) and into different isolated network segments.  Hopefully by the time I scale beyond a single deployment SolarWinds will have something better than EOC or a more functional EOC.

Level 12

I can't be there adatole​ but thank you for this

Level 16

hope for news from that FR

I quite sure that every poller  can handel much more 12,000 elements

well it's politics...


adatole​ I look forward to finally getting to meet you at ThwackCamp. I'll be onsite at Austin. I have no idea what I am going to say because I am lousy with smalltalk. lol! But............   Yea!  😉

Level 18

No need to feel bashful or shy, we'll have plenty of non-smalltalk topics to talk about - the latest SolarWinds features, your environment, your burning questions, and more!

Level 18

Well hopefully you'll be able to join us online. But even so, the videos of all the sessions will be posted immediately after THWACKcamp, so you can "attend" later if pesky things like work get in the way. And as always, if you have questions about what you see and hear, toss them out on THWACK and we'll be ready.

I wouldn't say shy or bashful, per se. Just a general awkwardness until I get comfortable with my surroundings. lol!

Plus, I must confess, I've enjoyed all of your non-Solarwinds blogs/articles/entries/etc. most. I do enjoy your Solarwinds stuff too. But... you know what I mean. Like what you put last December. Good stuff...

Level 12

We really need best practices for all things IT.

Level 10

Is there documentation out there for the topics discussed in this session? I would like to implement some of their recommendations.

Level 13

I am also interested in any documentation.  I seem to recall it mentioned in the video that there would be at some point.

Level 10

adatolepatrick.hubbardKMSigma​ do you guys have these guidelines published?

Level 12

patrick.hubbardadatole KMSigma.  I have the same question as j_dennis.  Where can we find the guidelines mentioned in the video? 

Community Manager
Community Manager

I know that no one in IT likes to hear it, but the answer is: It Depends.

I'm sorry for being vague, but there are many reasons to scale this stuff out.  For the first, if you have "speed of light" issues, then you should use an Additional Polling Engine.  That's the "easy" answer.  Another great time to use one is when you have overloaded your Main Polling Engine (the first server you installed).  The last good reason is when you don't want to deal with obsessing over firewall ports.  Much polling (at least on Windows environments) uses WMI, and WMI uses a ton of ports.  If you use an Additional Polling Engine, you only need a few ports opened (1433 for SQL, 17777 for SWIS, maybe others - from memory).  Some of this has been alleviated with the optional agent, but that's not always practical.

Regarding Additional Web Sites - my personal recommendation is to always use them.  Even if you don't have over 25 simultaneous web users, using the web engine to reduce the load on the Main Polling Engine is always beneficial.  I spoke about some of the scaling concerns in a talk at our SolarWinds User Group.

Solving for Scalability - Kevin Sparenberg Chicago SWUG 2018 - YouTube

Please feel free to ask me any questions directly.