The software-driven economy is putting the heat on data logging, metrics and tracing technologies. Traditional tools of this type aren’t quite cutting it for application-first, digital companies. A bit of downtime or poor performance for a mostly brick-and-mortar business with an app is manageable. If a company’s application is its business, they could drive users into the up, available, live-streaming embrace of competitors.

“For companies for whom the code is the company, downtime is money,” said Christine Heckart (pictured), chief executive officer of Scalyr Inc., a company founded in 2011 that offers a unified observability and log management platform for application development and deployment.

That doesn’t only mean the potential loss of end-user business. It’s also true that the time and talent it takes to get apps up and running again is money. Software developers spend 35 to 50 percent of their time validating and debugging software, according to research from the Association for Computing Machinery Inc.

“The cost of debugging, testing and verification is estimated to account for 50 to 75 percent of the total budget of software-development projects, amounting to more than $100 billion annually,” according to ACM’s report.

Companies can save time, money, and their staffers’ sanity with superior debugging technology, like logging and metrics tools. There are some good, tried and true logging solutions out there for companies running traditional information technology environments and traditional apps, according to Heckart. For example, there is Splunk Inc. “They’re amazing at doing that,” she said.

But most log queries in traditional environments take several minutes. For application-first companies, that time must come down — way down, she added.

Heckart spoke with John Furrier (@furrier), host of theCUBE, SiliconANGLE Media’s mobile livestreaming studio, for a CUBE Conversation at theCUBE’s studio in Palo Alto, California. They discussed why improved logging and observability are critical in application-first and software-driven business.

This week, theCUBE spotlights Scalyr in our Startup of the Week feature.

Forged in Google’s bubbling code cauldron

Steve Newman, the man behind the software we all know as Google Docs, founded Scalyr in 2011. As a senior engineer at Google LLC for  years, he learned the software-driven ropes the hard, hands-on way.

“If you are delivering software continuously, there is no room in that process for a manual QA cycle,” he said in an interview with Infostretch’s DTV.

Scalyr’s software as a service was born out of the need for logging and observability optimized for the new stack. These are stacks that generally run on cloud with a large swarm of containers (a virtualized method for running distributed apps) deployed at any time.

The software systems needed to run software reliably as a business are complex. Monolithic structures are broken into smaller, more swift parts that can be difficult to keep track of.

“If  you’re cloud-based, if you’re moving to serverless, if you’re on Kubernetes or some kind of container platform trying to do orchestration, any of that makes it faster and easier to build the service,” Heckart said.

Unfortunately, it also makes it harder to pick the buggy needle out of the haystack. Engineers need a sophisticated detector to speedily pinpoint the problem within a service.

Subsecond queries find needle in new stack

“Scalyr’s designed by engineers for engineers on modern stacks to help them figure out where that problem is and get it solved very quickly,” Heckart said.

Scalyr is a multi-tenancy system with massive parallel processing. It can ingest any type of data, including machine and log data and very small, hard-to-categorize data events.

“When you go to run a query, we throw every processor in our system at every query that comes in,” she said. This gets query times down to a number companies who live on code can manage.

“Ninety-six percent of all queries happen in less than a second,” Heckart added.

Scalyr’s dashboard and search features are approachable and point-and-click intuitive. Users can query in plain English; there’s no need to learn a specialized query language. “In the more traditional systems, you find there are only a few people who know how to use the product, because you have to learn the query language,” Heckart stated.

And unlike other logging and observability tools on the market, Scalyr does not rely on keyword indexing. The problem with keyword indexing is that it keeps querying stuck in the past, Heckart explained. “We’re the only architecture that doesn’t use keyword index in order to do that search. And that’s what makes us fast,” she said. “What we have is this amazing, multi-tenancy, columnar-based approach that gives you these advantages of fast, simple and affordable.”

Tracing joins the bug-kicking team

Scalyr grew its core customer base last year by 170 percent in revenue. It announced distributed tracing at Amazon Web Services Inc.’s re:Invent conference last year.

Tracing is an important new tool for monitoring and debugging distributed systems, according to Heckart. Application monitoring systems are typically built for the front end. When companies move to container environments, they need the same capability for the back end.

“If you’re born in logs like we are, doing distributed tracing — which links them together and gives you a picture systemically of what’s happening and how you link the events for a fuller picture — we’re kind-of uniquely good at that,” Heckart said.

Scalyr’s tracing feature comes out later this quarter.

Watch the entire video interview below, and be sure to check out more of SiliconANGLE’s and theCUBE’s CUBE Conversations.

Photo: SiliconANGLE