I don't think that's an entirely fair statement. Cacti is not 1/4 the tool NPM is on many levels. Core feature of many opensource products in terms of presentation of complex data and proper handling of SNMP data for alerting, even if it comes from a table, is crucial and currently broken in SW and superior elsewhere.
It doesn't necessarily indicate an overall superiority, just a specific one. The lack of Product's ability to address these issues in any sort of a reasonable fashion is a completely fair issue to pound on though. It's been more than long enough.
No one said anything about being fair! How am I suppose to vent and be objective at the same time? This "product" is absolute anti-greatness. Look, it's not rocket science. We want to collect some data... we want to report some data. This concept is not some strange esoteric data voodoo being invented by the Solarwinds development team. It's called Relational Database Design and SW's implementation is EPIC fail. "CPULoad_Daily"? Seriously? They modeled an entire table for that? "APM_ComponentStatus_Detail" and "CustomPollerStatistics_Detail". Why?! I think SW's idea of normalization was making the UI a little nicer looking.
As far as what NPM does well... I can't find one thing that NPM can perform that other solutions like AppManager, Tivoli, OpenNMS, ZENWorks, or even *HOSTMON* can't also provide. SW needs a wake up call...
I attempt to vent and be objective at the same time, it's a strange need I realize.
I agree that there are many "anti-greatness" "features" of the product.
I also totally agree that the vast majority of what anyone *might* want to do with an SNMP collector / correlation engine is either missing entirely or very broken in the product. I don't know enough about database design to knock SW's, but certainly many things in the DB are somewhat less than straightforward. Last I heard, they were working on fixing that to lay the groundwork for more changes.
APM and UnDP's actually need critical work, although I haven't tried out APM for a version or two, because it was useless when I tried initially.
UnDP's on the other hand, are a constant source of pain, agony, and frustration. The simplistic use-cases that it actually works for are too manual, and too limited to be *really* useful unless the data you're trying to pull is a single, simplistic OID. God forbid you have a MIB that defines a table, or you're stuck in a pile of crapplication software that has you cursing the state of Texas, and especially Austin, which is completely wrong, Austin deserves a medal for existing in Texas.
Honestly, random feature bashing is getting old. The real problem here is a management one, not a technological one. These features can be worked on, they just aren't.
For literally every single SW software product I own, I have at least 1 or 2 critical bugs in existing functionality and at least 3 features requests that have sat on the ignore pile for nearing 3 years. The former is shameful, the latter is unfortunate. I don't ask for frivolous features.
SW continues with the mushroom treatment, and acts like it's a substitute for action. Enough already.
I agree with you 100% - it's a shame that Orion does not have this ability. I asked Mark Wigans the same questions some months ago in a different post, never got any reply.
It seems Solarwinds is chasing whatever feature is cool today, and they're still lacking in my opinion important features in the core product that should have been developed years ago. One should be able to display multiple pollers and values in a custom chart with ease in a product like Orion that is no longer either small or cheap.
You are all missing the most important thing here. We will have Meru Wireless support in the next version. MERU WIRELESS SUPPORT, GUYS! IT'S WHAT WE'VE ALL BEEN WAITING FOR ALL THESE YEARS!!! This feature was obviously high priority and asked for by thousands of customers, none of which apparently posted on the forums...anywhere...but it's higher priority than multiple pollers in one graph, because nobody except extreme fringe cases would want to do something so wild and crazy!
I very much look forward to having Meru Wireless support. Thank goodness this feature is finally being put into Orion. I was beginning to lose sleep over the fact that I couldn't easily graph my Meru Wireless devices without making a bunch of custom UnDP templates. My nightmare is finally over. I hope you'll join me in the praises of the product manager who finally put this feature into Orion 10 and made it one of the highest priority features. I was going to post on the forums that I wanted this feature, but it must have been such high priority that no one else posted about it, asking for it publicly, and I assumed that SW knew that we were all clamoring for it. In fact, I was going to post years ago that I'd sell tickets to my own grandmother's brutal beating just to see this functionality in Orion. But now I don't have to do that because it's going to be there! YAY! Please rush this version out the door because I desperately need the high priority Meru Wireless feature set, just like all the other hundreds of non-posts asking for this high priority feature.
Have I gotten my point across? I like the SW products and for the most part, they do the job for us. Sure, there's some disappointments, but I've always overlooked them because SW administration saves me a lot of time when compared against Cacti or MRTG in our environment. However, it really ticks me off to be told that all these issues were the highest priority issues when I don't see one single post from anyone asking for Meru Wireless support in the next Orion. You have several threads asking for this feature, most notably this one, and it's ignored and "sent to the PM", which I've seen tons of times and rarely does any suggestion about core functionality actually make it into the product.
To both Neckmeats and Pserwe
Both of you sound as though you feel that the SW products are completely not meeting your needs? I am curious why you haven't moved to a different solution?
I ask because this feels as though it has become an Orion bashing session versus constructive feedback opportunity. I am guessing that maybe this is because you both like the product for the most part and are just extremely frustrated that it's missing a few key component to be a much greater product for you?
I am not trying to dismiss you frustrations as I completely agree with them, all of your points on this missing functionality is completely validated. I also came from products like OpenNMS and Cacti that use RRD for graphing and more dynamic reporting so I feel the pain of no longer having that functionality. I also agree that being able to pull your data together and display it in a graph that you can control is at the heart of the relational database concept.
I also don't necessarily agree that all of the new features they include are low hanging "marketing" fruit. The UnDP for example was a much needed enhancement, while not as useful as it could be if we had Multiple Datapoints in a Single Graph, it was still a necessary "core" feature. Many of the features I have requested have been added to the product or are in development (I have seen mock-ups already) which are the types of things I have had as "core" with other NMS systems such as OpenNMS, HP OpenView, Manage Engine, What's Up Gold, etc.
What I am trying to understand is how the rest of the product is working for you and if not why you are still using it?
Also, what would you guys think if SolarWinds not only let us put in feature requests (as they do now), but also implemented a system where each of their current maintenance customers had a pool of points they could use to put toward any feature on the list (once the feature was implemented you would get those spent points back). This would give us more of a voice on what is top priority for us. This is a system I have seen used in other places and thought it may be a good option for SolarWinds.
Just wanted to solicit your thoughts on some of this stuff, thanks guys!
2. Relatively performant considering alleged database shortcomings, which I'm not qualified to have a real opinion about, I don't know anything about database design, nor do I pretend to on TV.
3. There is a lot of functionality in the product, it's just missing some key things. Now, I don't care about blackout windows, but I could totally agree that it makes huge sense. I just turn shut down my MTA when I'm doing something spammy, and that keeps the phones and Outlook from blowing up. Good times.
4. The presentation is pretty good, overall. The web interface is pretty well designed, and for things that don't hammer the database, relatively performant despite being overtaxed, this gets better with v10 and massively better hardware. I'm excited.
5. The support team is really coming along. Used to be years ago, it was small and tight, and other than a rare exception here and there, calling into Austin was decent. Then it expanded, and even though I swear I should get coverage at 2pm on Friday afternoon, PDT, apparently no rollover last week. When I can contact them, after validating basic things, I get solutions to complex issues the product wasn't designed to do sometimes, and other times, issues resolved. Right now I have a couple that aren't resolved, but until I'm running 10, I may be stuck. I can accept that. There are a couple of guys in support who I owe some beers to, should I ever find myself in Austin The lower tiers use of the backline support and Tier 3 is really solid, much better than it used to be. Honestly, the support guys and girls are all pretty great. I could make honorable mention, but I'll skip it for now.
6. The promise of a really tightly integrated suite of products. I want NCM integration to be as tight as APM and NetFlow, just like another module. It's not there yet. I'd like to be able to set permissions and have them fully respected by all the modules, at least in 9.5.1, not there yet. I'd like more granular permissions in general, probably involving adding a column to the nodes and interfaces tables, pretty straightforward. Oh wait, I kind of do that with custom properties so far, but it's far from complete, and custom properties on interfaces can/will be ignored for things like syslog, traps, and Netflow. I don't want someone to see the parent node, just 3 individual interfaces.
Harder than you would think to accomplish. So many resources on my network summary home ignore node and interface limitations. Netflow is the worst offender.
7. Alerting based off polling is pretty decent. Coupled with a UnDP, if you're pulling status from an OID that returns a single value, between transforms and data types, it's pretty damn flexible. I'm not gonna lie, I'm pretty much satisfied with that functionality. The problems come in with tables, not being able to update my own MIBs, and any situation beyond single value OID response. Syslog/Trap alerting is all changing when I start using 10, I haven't tested each one individually, but I will before I move over to it. It's a business requirement to make sure the transition is smooth.
8. I see progress in the product. I hear rumors occasionally about something cool that is rumored to be on the list to develop sometime. I see the product of easily becoming much more than it is today, and being indispensable to my and many other organizations. I've heard Product people talk about DB changes in every release and a migration to an architecture that will allow some of these features I'm asking for, but I'm not seeing it happen in any sort of timeframe that helps me, so I'm a bit upset.
I guess those are the reasons I like the product at the moment.
I'd love the points system, as long as it's scaled based on something like licensed products owned, and what size. I don't think an NPM 100 customer should have the same voice as SLX or multiple SLX. The majority may still swing towards a different direction, but the process would be transparent, which is another beef I have with SW.
I am absolutely positive I'm not the biggest license holder around here, and I am not even considering resellers in that mix. Still, I'm confident that most of the things that are on my "Need to Have" list also have a place on the lists of others who use the product in a similar way that I do, especially at any volume, with diverse data collection and correlation needs. We're just in the early stages of starting to go off-board for things we simply can't do, but it's not really a road I want to go down.
I also greatly desire my administration overhead to go down. Every detail we need to know to provision nodes, interfaces, users, and custom properties is stored in a database. I need to be able to flow that information through. A configurable CSV import filter would work, but finishing the API up for being able to write values is high on the list. SOAP/xml it is? So be it. Need write capability
SolarWinds solutions are rooted in our deep connection to our user base in the THWACK® online community. More than 150,000 members are here to solve problems, share technology and best practices, and directly contribute to our product development process. Learn more today by joining now.