Transcript
I am the senior product manager over the Fleet Manager product, so I'm here to give a quick update on that. I was really happy with the intro music. That's the thing I was most nervous about is which song I'm gonna get, so... I am gonna try to make the most of our time here today, so I know, obviously, many people are using Fleet Manager, but unlike HyperCore, not all of you are, so I'm gonna try to start out with an introduction to what is Fleet Manager since it was launched two years ago into what we've done since we talked about it at Platform, and by we, I mean Craig, actually. This is my first Platform event. It's really amazing, isn't it? Last year, I wasn't able to make it because I had twins, and I really am kicking myself for not having my first slide be a picture of the now one-year-olds, but instead, you get this slide, so... Just a quick agenda. What is Fleet Manager? What's new? And the bulk of our time is gonna be spent on what is next. So what is Fleet Manager? Believe it or not, it was launched a little over two years ago now, and so I'm just gonna cover the first span of what was there at release. A lot of you were beta customers at that time. What came out in the next year? I'm gonna start off with sort of what is the philosophy of Fleet Manager. I'm not gonna be able to say it as eloquently as Craig or Scott might, but Fleet Manager is carrying on, like, the long tradition of scale computing, simplicity engineered, right? So there's a lot of remote monitoring and management solutions, but we're trying to do remote monitoring management the scale computing way. A very smart person you may have heard of once said that you should make things as simple as possible, but no simpler. What we're offering is a complex solution, but we're still trying to bring simplicity in every feature that we add to any of our products. So what makes Fleet Manager, you know, monitoring management the scale computing way? Well, this is gonna sound familiar. It's easy to set up, right? So if you have used Fleet Manager, you'll find often the setup time is quite literally zero. You can log in, and your clusters, if they're initialized, are already there, and you can even actually initialize clusters through Fleet Manager. It's easy to use, so we have a user guide. I'm not sure that anyone's looked at it, because you can just log in, and there's your clusters, there's your nodes, your VMs, conditions, it's all just sort of plug and play. And then lastly, you know, we're making it simple, but not... it's not a toy, right? So it's easy to scale. Right now, as we're talking, there's 14,000 plus and counting systems that are communicating in real time all the data you need to accomplish your remote monitoring and management. So that's the philosophy of Fleet Manager. A little bit more of the details. I'm not gonna give you, like, a full product demo here. If you want that, there are essays out in the expo hall later, or find me, and I'd be happy to do that. But a quick overview of, in this first year, what was Fleet Manager, right? Maybe the last time you saw it. So as Craig mentioned in the security presentation, right out of the gate we had single sign-on with Google, Microsoft, role-based access control so that as we were giving you this powerful, easy-to-set-up monitoring and management, we were not skimping on security at any point. The third bullet point here is interesting, and I do want to call it out, especially for our partners, our MSP partners, that Fleet Manager, already from day one, enables fully managing your clients' clusters, excuse me, and it also allows for co-management. So you can sell Fleet Manager to a customer org. So if you're an MSP, I've got five clusters, maybe two each, so I've sold them a production and a DR node, and now they have that, they have Fleet Manager, really, I'm providing the day-to-day service to ensure that those are running and those workloads are performing optimally. They can invite you into their Fleet Manager org. So it's a multi-tenant cloud application, so you can sign in to Fleet Manager but then just toggle between different organizations. That's a neat feature I wanted to call out. And of course, once you're in, auto-syncing, reliable real-time data, that last, zero-touch provisioning, I'm going to fly over that because I know we spent a lot of time talking about it last year, but that's really cool. You can take the nodes that we ship to you and you can provision them straight from Fleet Manager, so no keyboard, no monitor, no having to be the guy on-site even, just fill out a form in Fleet Manager. We're going to be carrying on that tradition of zero-touch provisioning with some of the new features I'm going to talk about later in the presentation. And then, as Craig mentioned, HyperCore update management. HyperCore, we're kind of getting to steal some of their thunder, right, because they've made updates really, really easy, and then we get to kind of come along and make them even that next step easier. You can have 10 clusters you want to update, you log into Fleet Manager, it's a single click from Fleet Manager, so click, click, click, and then you can just watch the updates from Fleet Manager, really, really simplifying your day-to-day management tasks for your whole fleet. So, that's the history lesson, right? All right, good, close the history book. We're going to talk about what's new. So, a lot of you maybe are like me, and you weren't here last year, but I do want to quickly fly through what was presented last year and just demonstrate that the things that we're adding to Fleet Manager aren't just, you know, the things that pop into our head, they're based on the feedback that we get from you in calls, in conversations, in events like this. So, the first thing that we were talking about last year that has since been released is the VMs list. So, a list view for your VMs has always been a popular request. Fleet Manager now presents that to you. So, if you have 100 clusters, you can search all of those clusters in a table view, search, sort, and find the VM you're looking for, see patterns. But you can also enjoy that list view, you know, even if you're just looking at a single cluster. On a very similar note, current conditions, there's a power in aggregating all of the conditions, right? So, when you log into your Fleet Manager dashboard today, you get to see every cluster in real-time reporting up any conditions that may be occurring, so that instantly, without sorting through your inbox or some other tool, or logging into each HyperCore cluster, you can say, this is the most critical condition on my whole fleet, this is what I should be doing, you know, when I start my day, or, you know, whoever's responsible for that. You can also use it to say, I know about that condition, I've got a case open for it. Let's see what happens since I last logged in. So, you can sort by set date or search for, is a given condition present anywhere in my fleet? And then, of course, Craig talks quite a bit about SecureLink, but what I really, really love about SecureLink and what makes me really proud to talk about it, is that this was not our idea. This was our partner's idea. So, fleet management was about, you know, the monitoring and the reporting, and, you know, the stats, but they were like, hey, you know, what I really need is an easier remote access tool. And, you know, the light bulb went off, and we were able to deliver that this last year, and now this is really sort of the crown jewel of fleet manager. I highly recommend, if you have not checked it out, go and, it's a bit of a dopamine hit to be able to just remote access your clusters securely with a single click. The fun story that Craig almost stole my thunder on was, last night at the rooftop party, I got to run into Stephen Henshaw from In-N-Out, and, you know, we were having drinks, and he was, obviously, he saw me, and, oh, fleet manager. And what he did, is he pulled out his phone, he's like, you know, you can, I love being able to pull out my phone and go into fleet manager, and I can go to the cluster right here. And it was so great to have, you know, a customer, a partner really have that level of excitement, they're like, look at this thing you made, isn't it awesome? Yeah, I really, I mean, that gives me a lot of joy. Obviously, SecureLink, it adds productivity, as Craig mentioned, you can cut out VPN clients, or at least reduce your reliance on them, as well as jump servers, so there's cost savings associated, and there really is security benefits as well. You don't have to be responsible for configuring and maintaining those. We handle all of that for you. The bullet points really cover what Craig mentioned. So, access to the UI is still behind RBAC, it's still, it's all encrypted, it's only transmitting the HTML of the UI. All right, so, what's next? I'm gonna try to slow down, because there's some exciting stuff in here. I think it'll be exciting both for people who are currently using Fleet Manager, as well as those who aren't, they see themselves as, oh, I don't have enough clusters, or I don't really need a monitoring tool. And that's fair, and we've been listening, and I'm really excited to talk about a few things that are coming in this next year. The first one is a Fleet Manager API. So, if you have not used the API on HyperCore today, that's been around for a while, I highly recommend you go do that, you can do anything on the cluster, really, through that API. What the Fleet Manager API will do for you is add an additional layer on top of, much like the Fleet Manager UI, this will be the API to the UI. So you can, that current conditions list, okay, it's cool that it's in Fleet Manager, I want to pull it into my service and ticket management system in real time. Now you can do that with the Fleet Manager API that will be coming. ServiceNow, ConnectWise, you can gather real-time and historical data, I'll be talking more about the historical data part of that coming up. You can integrate the known cluster data into asset inventory systems, or you can get live cluster inventory for Ansible automation, I know Dave Demlow was talking about that in his Ansible presentation yesterday, if you were able to catch that. This greatly simplifies the task of, okay, I've got all these Ansible scripts, where are they going? How do I run them? Fleet Manager can provide that up-to-date inventory for you. So this is something that is in development today, and if you are interested in getting beta access as soon as it is available, please talk to me at any time later today. The second feature I wanted to talk about was zero-touch, node, add, delete. So Craig teased this in the product roadmap presentation, and obviously he had a super cute video of his daughter, so I'm not going to try to match that, but I did want to bring it up because this is a really exciting feature coming to Fleet Manager. It's building on the vision that we started with zero-touch provisioning, but you only provision once. You might add, remove, replace a node a couple times in a node's lifetime, especially when you get to the refresh. This could handle an entire refresh really without workload interruption. So in case you missed the product roadmap, I'll just fly through these. You get into Fleet Manager, you see, okay, my workloads are all still fine, but one of my nodes is down. So I'm going to need to get a new node out there, swap it. I don't want to have to do that. I don't want to have to bring my keyboard and my monitor out there to accomplish that. Call into skill computing support. So now you can hit that edit button. You can see, all right, this third node, that one's got to go. I'm going to click replace. I'm going to select the new node, and one of the nice things about this, and it's much like zero-touch provisioning, is this doesn't have to be done in the moment. You don't have to be on the phone to do that. You could, as soon as you've shipped this node out, you could come into Fleet Manager, say, hey, I just shipped this node, so I'm going to set up my config so that whenever it arrives, they don't even need to call me. The config is here in Fleet Manager. All I'm going to have to do is, that user's going to have to follow a little, like, four-step unplug this, plug this in, and the config that I've specified here will be delivered as soon as possible. There's not going to need to be any interaction from me after I've set up this configuration. And so, once it's plugged in, once it gets internet, you will then end up back with a three-node cluster as though nothing ever happened. So, really excited about that. The third thing, there's four, the third thing here is really radical. As I was preparing for this presentation, my main concern was that I was going to really under-emphasize the importance of this feature and what this is going to do not just for Fleet Manager, but for the platform, for the way that applications get onto HyperCore. So, zero-touch node add remove, all that node lifecycle managing the cluster, but the cluster is there to run applications. The application's king, so what we are working on is zero-touch application deployment so that from Fleet Manager, you can specify not just what are the nodes, but what are the applications? That's what matters. So, you can specify I want these three VMs, here are the specs, the RAM, the storage, here are the images, so this one's a Linux VM, this one's a Windows IoT VM, whatever. You can specify the number of HyperCore clusters you want as those targets and each cluster, again, as soon as it's online, so you're not waiting around, you're not clicking retry, this will be all kind of set and forget. Each cluster will remotely retrieve the configuration you specified, pull in any images, so you don't have to stage those images locally on the cluster, and then you can also leverage CloudInit and Ansible like you can do today, actually, on the cluster, but this automated system from Fleet Manager will also be able to leverage CloudInit for configuring a user in those VMs, Ansible for setting up the app. I'm excited to get out of the slides and actually show a demo of this functionality right now. And I'll just talk through, so on the left, obviously, we've got a HyperCore cluster, on the right, we have a ZeroTouch app deployment manifest, so that has the specs for the VM, it has the specs for CloudInit to set up a user, and also we're going to install Nginx and configure it so that we can deploy a warehouse application, and we've set up the warehouse application that we want, you know, could be weeks in the future, in the past. Now the cluster's up, it's going to pull down the image that it's needing, it's going to create the VM, it's going to boot the VM, the VM will get its IP assignment, and that will allow Ansible to run, to set up the configuration, you can see here, you can just grab the IP of that VM, and we can confirm that it is now running. Pretty cool, right? And so, you were watching it happen, but the real magic is, you could set that up one time, and then have that apply across, you know, sure one cluster, but also a hundred clusters, a thousand clusters. So, and that's leading us towards that container story that Craig talked about as well. The next thing, the next and final thing I wanted to share with you all is historical resource metrics. So this one, as I mentioned, SecureLink was born from your feedback to us. This one, I think it's safe to say, has been around for a long time, and we're excited to finally deliver on it. So, you can review up to two years of history for your CPU, memory, and storage of your HyperCore clusters in Fleet Manager. You know, I don't think I need to share the use cases with you, but of course, it's helpful when you're troubleshooting application issues. It's helpful for balancing the VMs across the nodes, particularly for CPU usage, so that your applications are achieving optimal performance, that you're making the best use of your resources. And then, of course, sizing and predicting node ads, or if you've got a refresh coming down the road, you can hop into Fleet Manager, see, what's my trendline here for storage? How is that going to help me size appropriately? So, walking through this, we've got mockups here of CPU, memory in the middle, storage on the right. By default, you're seeing the last hour, and that's really exciting. I hope for all of you, I'm excited, but we don't want to stop there, right? I mean, cluster metrics are nice, but if you see something, you see a spike, you see a valley that you don't understand, that is definitely wrong. What you are going to want to see is the node level metrics, of course, so that we're also working on that. So you can see, in this mockup, we've got one node that is kind of running full tilt on CPU, so easy fix, go into HyperCore, drag and drop, move the VM over to one of the less utilized nodes for CPU. Same thing for memory, you can get one view of clusters, down to the nodes, ensuring balance, ensuring that if they have gone out of balance, how did that happen, when did that happen, and with storage, what's nice, I hope everyone here knows that one of the benefits that I think kind of gets undersold because we all take it for granted at this point, because it just works, is that HyperCore will balance your storage across your nodes, of course. Scribe will do that for you, but we didn't want to leave this blank space on the bottom, so we're not going to stop at nodes, we're going to go down to the VM level. Again, the application is king, and we want to help you make sure that they are running appropriately. So you can see your CPU stats right alongside your node and cluster stats for optimal allocation. Same with memory, and same with storage. So I hope that that will help with troubleshooting, it will help with getting the most out of your investment in your skill computing clusters, both today and down the road as you're refreshing nodes, changing nodes. And I just showed a lot of slides, so what I want to do next is show a little live demo of cluster metrics. So this is what every presenter loves to do, is do something in production in front of hundreds of people. We've got to get out of the slides, and we've got to see that this is in Fleet Manager today. You can see your CPU, you can see RAM, and you can see storage over the last hour, over the last week, and when you see a spike or a dip, you can even zoom in. So right here, we see our capacity dropped. We can zoom in. Huh. On April 13th, this cluster lost capacity, and lo and behold, it exactly correlates with what HyperCore already knew, and was telling you with this condition the drive has failed. So you can see, really at a glance, almost intuitively, the impact that's having on your space left. And talking about that node add refresh use case, historical metrics are more useful the more you have, right? And so what's exciting is, we didn't start today, we actually started back in September collecting these metrics. So when the feature's available, you'll have that to leverage. So that's the live demo, and what I'm really excited about is seeing since it is live, I really wanted to share it. So if you are a fleet manager customer and cluster metrics looks interesting, useful for you, it's something that you've been waiting for for a while, or just you want it today, please let me know. We're excited to start giving beta access to that today. If you're not yet a fleet manager customer, I'm not a good sales guy, but what people are often surprised to hear is fleet manager starts at just $400 a year for your first five clusters. So it's not an upcharge, it's that fleet manager is part of the platform that we're trying to deliver to make sure that everything is simple, scalable, easy for you. And I hope that's what you saw today. If you have any questions, you want to see something more about a feature, you want to let me know what the next thing next year platform, what you want me to be presenting is, this is what we're working on, please find me. And thank you so, so much for your time. This has been a blast. Thank you.
