 Alive from San Francisco, it's theCUBE. Covering AWS Summit 2017, brought to you by Amazon Web Services. Hi, welcome back to theCUBE. We are live in San Francisco at the AWS Summit. We've had a great day so far. And Lisa Martin here with my co-host, George Gilbert. We are very excited to be joined by Datadog. Kay Young, the Director of Strategic Alliances from Datadog, welcome to theCUBE. Thank you, hi, glad to be here. So tell us, besides loving your shirt, as I've already told you, tell us and our viewers a little bit about who Datadog is and what do you do? All right, so Datadog does infrastructure monitoring and application performance monitoring. So what that means is we're able to not only look at your hosts and the resources they have available to them, CPU and memory and that sort of thing, but also all the software that's running on top of it. So if it's off the shelf software, like a database, like Postgres, or maybe it's Engine X, we understand over 200 different off the shelf types of software, integrate with them directly. So all you have to do is turn on those integrations and we can tell you whether those pieces of software are performing at the rate that they ought to with a sufficiently low number of errors. That's the infrastructure monitoring side of things. Then application performance monitoring is where you can actually trace execution of requests, individual requests across different services or microservices and tell where time is being spent and track metadata so that in a forensic case you can go back and determine, oh, this type of call is producing a lot of errors. Oh, and those errors are coming from here and then maybe a lot of time is being spent here and then because Datadog also does infrastructure monitoring, drill down into, okay, well what's happening under the hood? Maybe we're having problems because our infrastructure itself is misbehaving in some way. You've some pretty big customers. Salesforce, Airbnb, Samsung. I was just reading yesterday an article that was published that you've been, Datadog, in the top five businesses profiled by IDC as the multi-cloud management vendors to look out for some pretty big accolades, some pretty big customers. How long have you been in business? Since 2010? 2010 and tell us about what you're doing with Amazon. What we're doing with Amazon. So let's see, where to begin. Amazon, a lot of people come to Datadog when they have complex systems to manage, meaning highly dynamic or high-scale or they've adopted Docker and their infrastructure is changing frequently, more frequently than infrastructure used to change 10 years ago because Datadog makes it easy or easy, possible even, to make sense of what's happening even as your infrastructure changes on an hourly basis. So a lot of customers come to us around the time they're interested in using dynamic infrastructure. Sometimes that's on Amazon and sometimes that's when you're on-prem but you're adopting Docker, for example, or microservices. We get a lot of business on Amazon. Amazon, I think it's fair to say Amazon loves us because it makes it so much easier to use their service and to adopt their service and we're sort of the de facto infrastructure monitoring service for Amazon. So you talking about containers, microservices, hyperscale, is there a break with earlier monitoring and management software that didn't handle the sort of ephemeral nature of applications and infrastructure, is that the change? Yeah, that's basically it. 10 years ago, you as an assistant administrator and operations person would have known the names of every one of your servers and you kind of treat them affectionately, oh, you know, old Rogers misbehaving again, we got to give it a reboot. These days you don't know, in many cases, even how many servers you have, much less what's running on them. So it used to be that you could set up monitoring where you'd say, okay, I need to look at these things, they should be doing these set of tasks and you'd set it up and you'd basically forget it for six months or a year. Now what's happening on any given machine or inside of a container is churning very, very frequently. And so to make sense of that, you have to use tags. So you tag all of your infrastructure with what it's doing, maybe what environment it is, like if it's staging or production, whether it's in AWS or on-prem, maybe it's a part of a build. And then you can look at your infrastructure and its performance through those lenses. You don't have to think in advance, oh, I'm going to want to know what's happening in US East One in production with build number 1180. You can just do that on the fly with data. And that's the sort of thing that we make possible that's necessary for modern applications and modern services that really wasn't possible before. So it sounds like it's fairly straightforward at the infrastructure level to know what metrics and events you want to collect in the sense that CPU utilization, memory utilization, and maybe even a database number of connections and query time. But as you move up at the application level, the things that you want to ask could become very different between apps and then very different across cloud or on-prem. Yeah, that's right. So there's sort of two classes of different things you could want to ask. Datadog accepts totally custom metrics. So we know about, as I said, 200 different technologies and we can collect everything automatically. But then you're going to have your own application and you're going to want to send us things that are specific to your business. We take those just as well. So for example, I think we have one customer who tracks when cash register drawers open or close. You know, that's not built in, but they can send those metrics to us. They get graphed the same way. We can set alerts on it the same way. We can use sophisticated machine learning to make projections about how we expect those patterns to be in the future. And if the cash registers don't open at the right rate, we can let somebody know that something's gone wrong. So we can collect any kind of metrics. And then on top of that, we've got application performance monitoring, right? So that's where you've written custom code and Datadog, since it's already running on all of your servers, can track requests as it moves from service to service or between microservices and recompile that request into a visualization that'll show you everything that happened, how long it took and allows you to drill in and get metadata about each thing. So you can actually reconstruct where time is going or whether there are problems. Why don't I ask you about some of the trends? As I mentioned a minute ago, reading that article or the mention of Datadog by IDC is one of the top five multi-cloud management vendors. What are some of the trends that you're seeing with respect to hyper-cloud, multi-cloud? We've heard some conversation today from AWS, but I'd love to get your feedback on, as the director of strategic initiatives, what are you seeing? Yeah, so the trend that I'm going to answer is, the trend that we were seeing a few years ago was more and more people were adopting cloud, period. And that's continued and continued and continued. 18 months ago, if you went and talked to a large financial services organization and you told them, we do monitoring, okay, they're interested, well, we run only in the clouds. You actually have to send your data to the cloud. They'd show you the door very politely. And now they say, oh, well, we're going to the cloud now too. And so it's a great place to be. Now we're seeing organizations of all sizes of all types are in the cloud. So the next leading trend is containerization and microservices. So we actually published a Docker adoption report. We've done it three times now. We refreshed it yesterday. We do it about every six months. And we take a look at all the usage that we can see, because we have this somewhat unique vantage point of being able to see tens of thousands of customers usage, real usage of infrastructure, and look at, okay, which percent are using Docker? When they use it, do they dabble with it? Do they fully adopt it? Do they eventually abandon it? What are they running on it? So we've published a very long report. Anyone that's interested can actually Google Docker adoption and we'll be the top hit there. We've got eight different facts that talk about how quickly it's being adopted. Docker adoption is really quite remarkable. We're seeing a 40% growth in true adoption, not just dabbling since last year. At the same time, we've seen a more than 100% increase, a more than doubling of the companies that use Docker that are using orchestrators, like Kubernetes, to manage even more sophisticated and rapidly changing fleets of machines. That's really meaningful because orchestration with containers really enables microservices, which enables DevOps, which enables people to move quickly with very little friction and own specific parts of a stack. Does that mean that their on-prem operations are beginning to look more and more in terms of processes like the clouds? That it's not just a VM, but they're actually orchestrating things? Yes, it does. And people will run orchestration on top of the cloud or they'll run it on-prem. But yeah, it's exactly the same. It's the same idea if you're on-prem, you have a physical machine, you're running several containers in it and they can just be very fluid and dynamic. And then, how does machine learning, how do you fit machine learning into the, whether it's at the infrastructure level or at the application performance management level? Yeah. Do you run it and get a baseline of what's normal? Yeah, so there's some very deep math behind what we do. So we're able to project where metrics ought to be in the future, not across any number of different categories or tags that you give us. It's important that we do that very accurately because we don't want to have false positives in our alerts. Meaning we don't want to wake people up unnecessarily. We also don't want to have false negatives, meaning we don't want to not alert when we should have. So there's a lot of math that goes into that and we can take care of very complex periodicity even while trends are happening within metrics. And doing that at scale so that it happens in real time is a challenge, but one that we're very proud of our solution. So you've been able to really drive or derive some differentiation in the market. One of the things I was also reading was that a lot of the business, I mentioned some of those great brands, is in the U.S. and your CIO has been quite vocal about wanting to change that. What's happened in the last year maybe with big rounds of Funding Me Rays that's going to help you get more global as even Amazon was talking about expansion and geographies this morning. Well, so it's even been a while since we've raised money, year and a half now, I guess. But the company is doing so well. It's a great place to be. The company is doing so well that we're just able to expand our operations and look bigger and bigger. Our two founders are actually French or they were born in France at any rate. And so we have a Paris office and we're moving pretty aggressively into Europe now. Fantastic. One question on again the hybrid cloud migration whether it's on-prem to say Azure or on-prem to Azure and Amazon, would the use of Datadog make it easier for the customer to essentially run the same workloads on either of the clouds? So we see a lot of people using Datadog at the moment, coming to Datadog at the moment when they need to move from pure on-prem to maybe hybrid or maybe fully into the cloud because you can set up Datadog to look at both those environments and understand the performance characteristics and then move over bits of it into the cloud and make sure that nothing's falling apart and that everything is behaving exactly as you expect. And then how about for those who say, well, we want to be committed to two clouds because we don't want to be beholden. Do you help with that? Yeah, we don't help with literally data movement which is sometimes one of the challenges. But in managing sort of a pane of glass. Yes, exactly. It's all one pane of glass and you can take once metrics are in Datadog, it doesn't really matter where they came from. You can overlay requests per second or latency in Google's cloud right alongside latency that you're seeing in AWS on the same graph or next to each other where you can set alerts if they deviate too much from each other. So it's kind of an abstraction layer or at least a commonality that customers would be able to have those applications in different clouds from different providers and be able to see the performance of the application on the infrastructure. And so one last question for you is we're getting everybody to wrap here. You know, there's a lot of debate about hybrid cloud and there's reports that say in the next few years companies will have to be multi-cloud. Just look at the SNAP IPO filing from a couple months ago, big announcement, $2 billion over five years with Google and then revise that S1 filing to announce a billion dollar deal with Amazon. So I'm just curious, are you seeing that maybe with the enterprises like a SNAP more and more that by default, whether it's for redundancy of infrastructure operations, is that a trend that you're also seeing that you're quite well positioned to be able to facilitate? Yeah, we're definitely seeing, you know, it's clear that Amazon is in the commanding position, for sure, but we are definitely seeing more and more interest in actual action in other clouds as well. Fantastic. Well, we wish, we thank you first of all for being on the program today. Great, congratulations on the success that you've had with Amazon, with others, and with the market differentiation. Congrats on expanding globally as well, and we look forward to having you back on the program. Right, well thanks very much for having me. Excellent, so Kay Young, Director of Strategic Alliances from Datadog. On behalf of Kay, my co-host George Gilbert, I'm Lisa Martin. You're watching theCUBE live from the AWS Summit in San Francisco, but stick around because we're going to be right back.