SignalFx, a cloud application monitoring company designed to help customers visualize issues and work at Web scale, came out of stealth today revealing two previously unannounced funding rounds and launching their product publicly for the first time totaling $28.5 million.
The first round was in March 2013 for $8.5 million from Andreessen Horowitz and Ben Horowitz joined their board. The second round was this past January for $20 million led by Charles River Ventures with participation from Andreessen Horowitz.
The two founders have first-class Silicon Valley pedigrees with CEO Karthik Rau coming from Delphix and VMware and CTO Phillip Liu, who spent 4 years working at Facebook as a software architect.
It was there that he came up with the idea of a product that would work at web scale (which Facebook required of course), that could monitor billions or even trillions of transactions and surface the ones that mattered most.
“I worked on the infrastructure team at Facebook,” Liu told TechCrunch. “I built a lot of configuration monitoring tools.”
“Monitoring using traditional tools is no longer useful or viable in a distributed architecture,” he explained. Facebook started by using open source tools for this purpose, but soon found (as it often does), that these tools were not sufficient to work at Facebook scale. Liu helped build some tools while there and some of the insights he gained led him to join forces with Rau and start SignalFx in February of 2013.
SignalFX is a cloud-based technology that streams analytics in real time, giving immediate insight into a problem when it occurs, Rau said. “Yes collecting 3 trillion signals a day is important, but you also have to be processing in real time and outputting analytics immediately to identify trends and anomalies as they occur,” he said.
Looking back isn’t all that useful in a world where you need to be on top of everything, he explained. He likened it to credit card fraud. If a financial services company figured out fraud was occurring an hour after the credit card was stolen, it’s really too late. The damage has been done. It’s the same in application monitoring. You need to see what’s happening in real time.
And that’s what SignalFX claims is its secret sauce. “We built SignalFlow streaming analytics technology to process huge volumes and output results [very quickly]. We have customers processing tens of billions of signals a day,” Rau said.
What’s more, the customer can use this information as a big data analytics engine to answer questions like which customers are experiencing the highest latency, which are using the application most often and so forth.
Rau says the primary competition in this space to-date is open source tools, which Facebook proved couldn’t scale (at least three years ago when Liu used them) and companies building their own systems.
In case, like me, you think application performance monitoring vendors could be competition as well, Rau sees them as a source of data for his tool, another set of information to give developers insight, but not necessarily as a form of competition.