New Automox Device Explorer: Never Draft Complex Queries From Scratch Again

Episode Summary

In this episode, Emily Pace and Jack Miller from Automox unveil Device Explorer, a revolutionary tool designed for efficient device management across organizations. They explore its genesis, spotlighting the query composer's flexibility from simple to advanced searches, which tackles the significant challenges of manual filtering and boosts user experience through regular updates. They discuss standout features, such as the ability to save and clone queries, export data to CSV, and promise future enhancements like dynamic groups and broader reporting functions. The duo underscores the engineering team's dedication to meeting customer needs, prioritizing user experience, and overcoming hurdles in data management and scalability. Furthermore, they hint at upcoming improvements and invite users to experience the performance advantages of Device Explorer on the Automox website, highlighting the team's commitment to relentless refinement.

Episode Transcript

Peter Pflaster: Hello and welcome to the 5th or something like that episode of product talk on the Automox podcast network. It's where we talk about all things Automox product. You'll notice I've got a little bit of a different crowd with me today. Steph is not on the podcast. So we've got two guest stars. One of them is Emily Pace, a product manager on the team. And the other is Jack Miller, one of our senior software engineers. So.

Emily, why don't you give us a really quick intro and overview of yourself?

Emily Pace: Sure, thanks, Peter. Very excited to be guest appearing today. Emily Pace, I'm a Senior Product Manager here at Automox. I've been here a little over two years, and have worked on some really exciting features, including the dashboard redesign. So really excited about talking about the Device Explorer today.

Peter Pflaster: Great, thanks, Emily. And Jack, Why don't you tell us a little bit about yourself?

Jack Miller: Hi, thanks Peter. I'm really happy to be here. This is really exciting. They don't let us out of the caves very often in engineering. So yeah, I'm Jack Miller. I'm a senior software engineer here at Automox. I've been here for quite a few years now, like two and a half, three. And I've had the pleasure of working some of our like flagship features like AVR. If you've seen our revamped login screen, that was me too. I've led some of the engineering efforts on some projects like Ask Otto, and most recently Dynamic Groups  or the Device Explorer, really excited to talk about it.

Peter Pflaster: Spoiler alert for the episode today. We're going to be talking about Device Explorer coming out I think in the next couple of weeks here. We've got an internal access period, so I've had a chance to go in and mess around with it. It's a really cool future. We think you're all going to love it. And we're going to talk a little bit today with Emily and Jack about the process of how we built it and why. So Emily, to kick us off, why don't you tell us a little bit about device explore what it is and why we built it.

Emily Pace: Sure. So like Peter mentioned, the Device Explorer, we're going to be introducing a new way to interact with your devices within your organization. Device Explorer is going to be added next to the existing device view today. That's not going away. We're just enhancing the functionality. And it's going to enable you to be able to create searches that are as simple or as complex as you want using a new query composer. 

So when the Device Explorer comes out, we're going to have device query composer, and a results table. And why did we build this? Great question. So we know that searching and reporting and grouping, they're all major pain points within our tool, and we want to solve that. And now that we've kind of hit our cadence with delivering features, we're going to be able to do that.

And we're really trying to deliver iteratively. We want to deliver value in small iterative chunks that you can use quicker. That'll enhance your usability of the tool. And so this is just one of those features that we're going to be iterating, smally, and will give you this ability to have increased searching functionality. 

So currently, when you're searching, you have to manually filter devices each time you want to export or view a set of devices. With the Device Explorer, you're going to be able to filter sets of devices down more extensively while also saving those queries for later. So you're going to be able to save these complex queries that you're setting up, and you won't have to do it again. And that's going to be available within your organization so other members of your team can use those queries. And we're also allowing you to clone them.

So if you want to change a complex query and you don't want to create it from scratch, you can clone it and just tweak it bit by bit.

Peter Pflaster: Awesome. And I know for those of you, those of us that are old school, I guess I'm not, I'm not super included in this. I'm against Excel these days. I've fully converted to Google Sheets, but there is an export feature as well. I believe if you're one of those people that wants to kind of take everything outside of the console and slice and dice your query as well.

Emily Pace: Yeah, thanks for that. You'll be able to do a .CSV export of the table when you have a set of devices.

Peter Pflaster: So the Excel people are fulfilled. Good for them. That's awesome. I'm really excited for our customers to be able to get their hands on that. But I'm also really excited about kind of the iterative aspect that you're mentioning, right? I have a pretty short attention span. So getting something new in the console for myself, more selfishly, just on almost a weekly basis at this point, we're improving stuff.

And I'm really excited about the vision here, if you could talk a little bit about kind of the broader product vision related to Device Explorer.

Emily Pace: So yeah, there's a really powerful service behind this query builder. It's a new service that one of our principal engineers and architects created, and it's going to be utilized throughout the console. So this is the first iteration of that service. With the service, with the front-end aspect in the console that you're going to see with the Device Explorer, we want to get you comfortable, like I said, creating these queries. It's something we haven't done before. It's going to be new.

You're not going to be as comfortable with it as you are with the current filters today. So we want to give you a chance to create these queries and get comfortable before we implement future iterations. Future iterations are going to include dynamic groups or collections, for example. That's a really big project that we have a lot of excitement over and a lot of anticipation on. And within dynamic groups, you're going to be setting up these groups like you are these queries.

So it's going to be a flavor of this query composer. And so that's why we wanted to release this before that much larger scoped project and change of functionality. And so when we get to dynamic groups, that's going to open up a lot more functionality in and of itself and usability within the tool from projects like maintenance windows, executing policies based on specific collections of devices.

And then expanding reporting. We know reporting is a huge pain point as well for customers and being able to expand reporting capabilities based on these dynamic collections of devices is gonna be huge for us.

Jack Miller: Yeah, that's super cool. Like Emily, I couldn't have put it better, honestly. Like the whole time, we're basically leveling up the Automox platform, right? Like in engineering, we've been undergoing insane changes to our process and asking a lot from our engineers to get to a point where we can iterate quickly.

Peter Pflaster: And now we're gonna have a... No, go ahead, Jack.

Jack Miller: ship code faster because we are like Device Explorer, I think represents a huge paradigm shift in the way that we are like thinking about the data in our platform and how we design features for our platform, the architecture of our platform too. It's a really cool story of evolution and change. And we're really excited like about the future of this platform, right? I think, you know, this principal engineer, Emily refers to Paolo.

Big shout out to Paolo, my main man on this project. We were working together very closely to make sure that these services that we are developing help customers get access to their data in insane ways and robust ways, makes sense, right? And we're gonna talk more about it, but I just wanna put that out there. It's super exciting.

Peter Pflaster: And I think the reporting aspect is not one to be overlooked. There are a lot of other tools today that have some version of a dynamic group or dynamic device tags. But a lot of the time, that final connection to the reporting isn't made, which for a lot of people, especially those organizations where compliance is particularly important or proving the work that you did, which is probably pretty much every organization out there.

If you can't connect that kind of group data to the actual patching, configuration, automation data at the end of the day, it really is a lot less useful. That's what I'm most excited about for our customers when we get to that point. But I've been messing around with Device Explorer today and it is super easy to use and really powerful and fast. 

I'd love to get into more of an engineering focused discussion here, right? So obviously Emily's job is to identify that customer need and figure out the best way to solve the problems. Sometimes it's problems that, you know, customers might articulate the problem in one way and Emily goes and investigates and it's actually an entirely separate problem. So that's her job is to help kind of unearth that and then decide what the business needs to do from an engineering perspective. 

But then engineering comes in and they really serve the product which obviously serves our customer orgs. So, Jack and everyone else in our engineering team, it's their job to figure out how to actually build it. So let's talk a little bit about that.

Emily Pace: Before we start that, can I add one quick thing?

Jack Miller: Yeah.

Peter Pflaster: Of course!

Emily Pace: OK, sorry, Jack. Just wanted to mention, just to piggyback on what Peter said with reporting and other iterations of the project, just to lay it out there, it's very iterative. We don't know what that's going to look like today, but we want to do it right. We want to build this the right way. We want to make sure it's what everyone is expecting. And so we have a really great UX team, and together they and I will be talking to customers and showing designs and really getting user feedback. So we're really excited about that. And if that's of interest, you know, we'd love to talk to you wherever you are.

Peter Pflaster: if you're a customer listening to this and you want to get involved, reach out to your CSM. They'll be happy to connect you with the Jacks and the Emilys of the world to get things going.

Jack Miller: yeah, that iterative process is very rewarding in the longterm, but you end up seeing very like in the game development industry, we call it white box kind of stuff, right? Things that are not like, okay, here's a Device Explorer. Like I can query devices, but what does this really mean? Like it's okay, it's a different view for devices, but really the valuable part there is knowing that it's going to be used in other parts of the platform. We're vetting the first fundamental, very primitive steps into what is going to be very robust functionality, right?

Honestly, it's a super exciting part of engineering because we get to ship code faster and we just love it.

Peter Pflaster: Yeah, I think the other thing that Emily mentioned is standing this up. It involves a new service that's pretty much is going to be as close to ubiquitous as you can get across the platform. I'm curious what that looked like and what kind of challenges we ran into and had to solve for.

Jack Miller: Yeah, I mean, basically the engineering problem you're trying to solve is, you know, Automox has a million devices plus in the platform, right? Like, it's all segmented by organization, right? Like, a customer is going to have their group of a couple thousand devices or a couple hundred or whatever. But ultimately, the SaaS platform, we have to figure out how to users, how do we store that, how do users access it? 

And so one of the ways that… I won't take credit for it. Paolo, our principal engineer on the project, he's really great. Him and his team worked very hard to try to solve this problem elegantly. And it wasn't without its complexities and its challenges. And it's just like a lot of data. And not only that, like we're talking about one device is like, we have a million of those, right? But every device has like a lot of properties, right? Every device has software. 

Every device has details about like a CPU disk on it and how much memory it has. We have to basically say, all right, if I want all Windows family devices, or if I want all Windows devices with operating system versions 10.3 Pro, we have to be able to get that in a way that makes sense and quickly. And so our standard, if you're an engineer, relational databases and all that stuff makes sense. 

Postgres, MSSQL, they make sense to the average person. This doesn't cut it. This is a use case where most engineers are like, oh, my tools that I've relied on so often, they don't work at scale here. And it was hard because we had to do a lot of research. And what we ended up doing is we ended up going for this graph database. And this is getting really in the weeds, but I'll just kind of high-level overview it. AWS offers a service called Neptune. It's a managed graph database provider.

And instead of storing things in rows, like you would see in a spreadsheet, we actually store devices, device properties as vertices in a graph. And the relationships, they're all connected. So Windows would be a node, and they're all connected to all the Windows devices. And so that way, if we want to get all Windows devices, we just say, OK, here's the Windows property node.

What all devices is it connected to? And then we parse that data out, right? And that's how we get the data in very high level, very primitive terms. There's a lot more going on. How do you import that data? And how do you get it? The ingestion is almost harder than the query problem itself. But those are some of the challenges. And another challenge is how do you pull the data out? 

And we have this API that we're planning on releasing eventually in the future where you yourself can use it to run queries against your environment. You're not limited to the Device Explorer UI element. The query is very SQL-like. It's a SQL structured query language kind of thing. And we have documentation for it. And we plan on publicly exposing it in the future. But that's all I'll say about that for now. But yeah, it is a very hard problem to solve. And we have very smart engineers trying to do that. Device is the bedrock of our platform. We have to be able to do that.

Manage those entities in a sane way that's scalable. And maybe the graph doesn't serve all of our purposes. And we have some other data structures under the hood that are helping us facilitate the future extension of that. But building the pipeline to get things stored in a way and ingressed into our different managed services is a challenge and a half. And for a large scale SaaS ITOps provider like Automox, it's not an easy feat.

Peter Pflaster: The first thing I noticed when I built my first query, when I got access to the internal early accesses is how fast it was. I mean, I was changing properties, looking for specific software names, and I could see the list of devices just updating essentially in real time as I was clicking the different search criteria. And for those of you that are really interested in the search criteria, it's like, you know, highly configurable Boolean search criteria essentially. So you can create all kinds of crazy slices and dices to get your list down to what you want. But the speed to me was the first thing I noticed. I mean, there's no load time essentially from what I could tell.

Jack Miller: Yeah, and your mileage is going to maybe vary on things, but we're always trying to improve it in the test data that we work with. We test things under load, and we even saw some really great performance under some of our larger organizations. I think internally we have some automation test orgs where there's thousands of devices that we're querying against. And we've got caching, of course, and all that stuff under the hood. 

So if you like want to go back and you add a criteria and you want to delete it, that query, those device results are going to come right back. And it's going to be really cool. And When designing the experience for that, we wanted to prioritize like, staffy feel, right? Because we're showing off that this backend is lightning fast.

Peter Pflaster: Yeah, that's awesome. I mean, I think when you look at the evolution of the space, endpoint management, there's so much more emphasis today on scalability and speed and efficiency of work. And we've really taken that to heart. I think if you're an Automox user and you've been using us over the last six months even, you've probably noticed pretty measurable performance results to the entire console and just the snappiness of clicking onto a different screen even.

Emily, Jack, anything else that we should be discussing around Device Explorer and the engineering and product efforts around it that we haven't touched on yet?

Emily Pace: Only thing I can think of is just hitting again that this is iterative. We're going to be merging some enhancements to this specifically around attributes. We want to, we're going to have an extensive list of attributes this first go around, but we want to make it more expansive. And so we do have projects in our backlog to hit all of that.

So if there's a an attribute in there when this comes out that you want to see that's not in there, go ahead and connect with your CSM. We can talk about it, and it's more than likely something we plan on adding in the future.

Peter Pflaster: Yeah, and I mean, with any good automation approach, the actual executing of the task is pretty easy, right? The hardest part is figuring out and narrowing your subset so you're not being too broad with the automation that you apply. And this is kind of the bedrock, the foundation, that's going to help us to enable you all to do that in the future. I know our whole organization is really excited about it.

Kind of this first feature, but also the iterative process. We're really invested in it from Jason, our SVP of product slash CISO. You've probably heard his podcast, The Patch Tuesday, and I think CISO series as well. But he's really invested in it all the way down throughout the engineering and product UX org. So really looking forward to the future there. I know that we're going to be putting up a more general description on kind of our goals as well on the public roadmap on our automox.com website. 

So definitely check that out as well if you're hungry for more information. But yeah, with that, really appreciate Emily and Jack, both of you coming on and spending some time talking about Device Explorer. So excited for our customers to get their hands on it in the next month or so. And look forward to having you back on as we continue to iterate here.

Jack Miller: Awesome. Thanks, Peter.

Emily Pace: Yeah, thanks so much.

Peter Pflaster: Thank you. Thanks all for listening again. We'll see you next month.