An Interview with Amit Agarwal, Chief Product Officer, Datadog
For many of us, infrastructure and application performance monitoring is something of a mystery. Performance monitoring in today’s dynamic, large scale, virtual, and cloud environment sometimes seems like black magic. It is certainly at least very challenging to track so many moving parts. Datadog is trying to change all of that.
I had the opportunity for a long chat with Amit Agarwal who not only shared information on Datadog’s approach to performance monitoring, but also how he sees the future of cloud computing, devops, and performance monitoring.
HostAdvice: One of the things I noticed on your LinkedIn profile is that you say how over the past 15 years, you’ve run every part of a software business. Tell me about that.
I started as a software developer at IBM and then at a medical imaging company. I then moved into product management. At first I was on the more technical side, but then I shifted to general management.
Now at Datadog, I am involved in everything: sales, marketing, and product development. I still stay very close to the technology even though I am no longer writing code.
HostAdvice: Please give us the 30,000 foot explanation of your product.
Datadog is a SaaS (Software as a Service) based infrastructure and application performance monitoring platform. What we do is seamlessly aggregate metrics and events across the full infrastructure and application stack and allow you to see them all together from a single point of view.
HostAdvice: I know that most infrastructure products expose metrics and events, but how do you get that information from custom applications?
We actually offer two options for applications to expose their data to us. The first option is a RESTful API and the second is through an open-source agent which can be embedded in the application.
HostAdvice: How do you define your market? Who is your specific target audience within that market?
The monitoring space is actually very mature. We are focusing on customers who run their applications in cloud environments – meaning they are dynamically expanding and contracting the servers and containers used by their applications. This is different from traditional monitoring, that assumes a static configuration of servers. Traditional monitoring tools don’t really work well in a world where applications are deployed in an ever-changing cloud/server/container environment.
HostAdvice: I have to admit, that despite being a technical person, the amounts and types of data gathered and displayed in your tool’s dashboard seems pretty complex and overwhelming. How long does it typically take an organization to configure and then feel comfortable with your toolset?
You are correct that the scale and scope of what we monitor is very deep and broad. But you can easily choose the data that is of interest to you by creating custom dashboards. Eventually, users and teams collaborate to correlate the different perspectives and data in order to figure out what is going on.
As to how long it takes to configure, a unique aspect of our system is that there is no need to configure servers other than to deploy our agent and configure Datadog. Even with 10,000 servers, you could do this in a few hours. We’ve developed some simple scripts to help with this process. There is no need to purchase professional services from us – our product is SaaS-based and works out of the box.
In terms of getting up to speed on using the system, the learning curve is fairly short – it is typically days to weeks, not months. One of the big benefits of our system is that since each user has his/her own dashboard, they can easily create their own custom views without having to go through the hassle of opening a support ticket, waiting, etc.
HostAdvice: On the other hand, I was struck by the seemingly very reasonable pricing structure for such a complex product. How do you manage that?
We wanted to provide a simple, transparent pricing model. This allows our customers to deploy more widely across all of their virtual machines and also allows us to get insight and to better understand customer usage. We have a special customer success team that will help customers if/when necessary (at no charge).
HostAdvice: How many active customers do you have today? Where are they mainly located?
We currently have over 4,000 paying customers. They are located worldwide, with the highest concentration in the US and then in Europe, Asia, and Israel.
HostAdvice: How would you describe your current typical customer?
Our customers are normally companies that have developed their own custom applications, but we have customers that run the whole gamut – from Fortune 500 companies to SaaS companies and startups.
HostAdvice: Who are some of your biggest customers?
We have a large telecom customer that is monitoring 20,000 server instances with Datadog and a large eCommerce company monitoring 10,000 server instances.
A few companies that we can name include Amtrak, EA, PagerDuty, Salesforce, Zendesk, Atlassian, and Netflix.
HostAdvice: Who do you see as your main competitors?
Initially, when we were just doing infrastructure monitoring, our biggest competitor was home-grown monitoring. However, with today’s multi-tier, micro-service architectures, it is hard for ad-hoc tools to bring all of the data together in one place. When you do it yourself, you just can’t get to the same level of agility and performance.
As we move into application performance monitoring, we have three main competitors:
- New Relic
- AppDynamics
- Dynatrace
HostAdvice: How do you see your tools as different and/or better than theirs?
What sets us apart is that we focus on monitoring very dynamic environments, including infrastructure, applications, and public/private clouds. You just can’t do that with traditional monitoring tools.
HostAdvice: How do you see the devops tools market evolving in the coming years?
I think the time is right for these types of products. Much more of devops is being automated and we see more and more management of platforms being done via code and scripts. We are also seeing more data-driven approaches to platform monitoring. The tools today are getting much smarter than they were 10 years ago.
HostAdvice: Do you have any particular thoughts about the future of the cloud computing market?
We are seeing rapid acceleration of companies moving large parts of their computing workload to public and/or private cloud platforms. They are not yet standardizing on a single platform, so we are seeing a lot of hybrid clouds.
HostAdvice: What is the story behind the name Datadog?
It was originally just a “placeholder” when the company first started. At the founders’ previous company they stored all of their most critical data on a server called ”Datadog.” With time, the name just stuck for us and had the advantage of sticking out in the world of standard enterprise software names.
HostAdvice: When I saw that your website was located at datadogHQ.com, I assumed it was because you couldn’t get datadog.com. But it turns out that you do own datadog.com and it redirects to datadogHQ.com. Why isn’t your website on a URL that exactly matches your company name?
When the company first started we couldn’t get datadog.com, so we went with datadogHQ.com. After a few years, when we were more successful, we did acquire Datadog.com. But by that time, we felt that changing everything was too much work with too little reward. We do use both domains for emails and some other things.
HostAdvice: Another thing I noticed was that you have a blog in Japanese in addition to your blog in English. Why Japanese? Is the market for your product so great in Japan?
Japan is a very interesting market for us. The way we have evolved, all of our sales and marketing is based in the US, but we have customers all over the world. We didn’t specifically target Japan, but we developed a very loyal following there and so we are responding to the demand for a blog and documentation in Japanese.
HostAdvice: What are your future plans for Datadog?
We are expanding rapidly in the application performance space – you will see a lot of announcements about new additions to our platform in the next few months.
Our goal for the next few years is to give devops engineers everything they need to troubleshoot from a “single pane of glass,” so we will be developing more components to enable that.
HostAdvice: How many employees do you have today? Where are they located?
We currently have about 300 employees, with the vast majority of them in the US. We have offices in Manhattan, Boston, and Baltimore and a new small R&D facility in Paris.
HostAdvice: How many hours a day do you normally work? What do you like to do when you are not working?
I don’t really have hard start and stop times. My working day is very fluid – it starts when it starts and it ends when it ends.
I know it is very geeky, but I love spending my free time playing Civilization.
If you want to get the best cloud hosting, click here and get the suitable one.
