Show Notes

Rajesh Raman (Twitter, LinkedIn) is the Chief Architect for SignalFx. What does that mean? In this episode, Rajesh and Ian dive into his role and how he focuses on not just the implementation of systems and processes at SignalFx, but also the best strategies and practices that will be needed months down the line. They also go into Rajesh’s history at Google and Facebook, and discuss cloud technology and the future of the industry.

Topics Discussed: Cloud platforms, app development, innovation, coding, monitoring, observability, microservices.

Introducing Rajesh — (1:15)

• Rajesh’s history with computers dates back to seventh grade and he taught himself programming on a computer that is now in the Computer Museum.
• Grew up in Bangalore, India, which is now considered the Silicon Valley of India.
• Has a Ph.D. in computer science and always had a sense that no knowledge was wasted.

What Rajesh is working on a SignalFx — (4:43)

• SignalFx’s CMO Tom Butta is also sitting in on the interview.
• In most other companies, the duties of the chief architect might be folded into the job of the CTO.
• The chief architect strategically looks at the technology roadmap to determine what the company should be working on six or 12 months down the line.

Rajesh’s role with customers — (8:12)

• Primarily the CTO is focused on the external-facing duties while the chief architect maintains a focus on internal functions.
• Tom: “The fact that we have someone whose sole focus is on the company’s architecture has enabled us to stay ahead.”

Why don’t more companies align the way SignalFx does? — (10:20)

• It’s kind of like making a movie and the role of a producer. Some producers are much more engaged in the way things get done, others sit back a little bit.
• Because SignalFx is focused so much on engineering, for them it makes sense to have engineers more involved in certain aspects of the business.

What SignalFx does — (12:25)

• SignalFx is a “cloud environment visibility platform.”
• Back in the day when cloud technology was relatively new, specific businesses had to build up the expertise to deploy applications at scale, to build them to operate them. Now, in the age of public clouds, any large application, system or company with a public strategy needs to do that, too.
• “People have been monitoring systems for as long as systems have been built. What has changed is that cloud environments make monitoring more complicated because they are larger in scale and scope.”
• What SignalFx has built is a platform that gives real-time insight into how your platforms are operating.

Who SignalFx works with —(15:35)

• The companies SignalFx works with spans the length and breadth of the industry.
• “It’s not so much the company itself or what industry they’re in, but rather their maturity in their digital transformation.”
• In larger organizations, there are observability groups deploying SignalFx to other engineers in the organization and SignalFx offers tools to manage usage.

How observability differs from monitoring — (19:50)

• Monitoring is a function that has always been around.
• “When you’re monitoring a system you always want to know is it up or is it down? Monitoring is just trying to answer that question. But that’s only the first step.”
• You also need to know why things are up or down and how to quickly fix a problem if one occurs.
• “We monitor SignalFx using SignalFx. We’re one of our biggest users.”

Trends to watch for — (23:30)

• Services you have to monitor are becoming much more numerous and the lifetime of components continues to shrink.
• “The universe of monitorables and observables is growing very large.”
• Collecting data used to be the problem, now it’s all about managing data.
• “The future of this space is going to be proactively telling you where a problem might be occurring.”
• Microservices are also going to be important to observe and understand.
  • You take a very large problem and break it into small problems and solutions you can deploy individually. This gives you flexibility, but it also creates a much harder architecture to monitor.
• Head-based sampling is a random sampling of data that many companies are using.
• SignalFx, on the other hand, traces every single data point.
• SignalFx has a series of Ebooks and papers.

How is real-time cloud monitoring evolving? — (31:50)

• SignalFx is currently the only real-time monitoring system available.
• “We look at our dashboards kind of like the instruments on an airplane. Real-time systems are critical because when you make a change to a system, you want to see what is happening and changing.”
• If you’re having a flash sale that lasts five minutes, and in that five minutes you make $2 million dollars, if something goes wrong and you can’t figure out what it is for two minutes, you lose out on a real amount of money. Real-time monitoring helps to prevent that kind of problem.
• “The real-time nature of monitoring is becoming a critical capability in operating large environments.”
• There is going to be a lot more automation that is needed in this kind of real-time monitoring.

Rajesh’s past at Google and Facebook — (33:15)

• Rajesh was at Google in 2002 and Facebook in 2009 and worked as a software engineer.
  • • He was “employee number 400-ish” at Google.
• “At Google, things were really buzzing. They were just taking off.”
• The number of machines was exploding day by day as Google built its infrastructure.
• “I got my hands dirty building real systems.”
• Even though Rajesh worked with a lot of big names in the tech industry, he was always impressed with how open and accessible they were. There was also a culture where you could debate ideas to find what works best.

Lightning Round — (49:10)

• Slack is the most fun app.
• Rajesh is learning to play the guitar
• Advice: “Your job is to make sure the right decisions get made, but you might not need to make those decisions. You have to create an environment where the best decisions can come to light.”
• Patrick O’Brian’s Aubrey–Maturin series
• Dan Carlin’s Hardcore History