SRE Culture and the Right Approach are Keys to Successful Microservices

Cortex was built by engineers from Uber and Twilio to help your engineering team better understand and improve your services. Ganesh Datta, Co-Founder and CTO  at Cortex, tells us they’ve been working on Cortex for over a year and a half and are very much focused on how to help organizations build a culture of reliability and service quality. According to Datta, “A lot of this comes down to the sprawl of microservices a host of organizations have today.”

Datta continues, “And so as your organization grows, you end up building a lot of microservices and you start to lose track of the quality of those microservices.” Datta finishes that thought with, “I think Cortex comes in to say: let’s get a baseline understanding of where services are at, what is our service maturity, what is the service quality, what are we building, and what is out there? And I think Cortex is the first platform that can let you do all that in a single place.”

Nikhil Unni, Co-Founder and Chief Architect  of Cortex, addresses the importance of culture in Site Reliability Engineering (SRE) when he says, “I would say that’s the primary problem we’re trying to solve. Over the last five or six years, with the proliferation of microservices, teams are kind of independently building out their services. But after many years of building new services, what happens to the role of quality? How do you get a rein in that and say this is what standard microservice should look like within our organization? How do you actually track that over time?”

Datta believes culture is something many organizations don’t think about when it comes to engineering challenges as a whole. He says, “For example, you have an SRT team, whose main charter is to help the organization build reliable software. You also have a security team whose main charter is to make sure that teams are building with security in mind. And then you have the actual development team, whose main charter is to actually build the features that the business cares about. And so each of these different organizations has their own charters competing priorities.”

Shifting to common SRE trends, Datta made a point to highlight modernization when he said, “A lot of organizations have been building software for a long time now, you know; Java virtual machines (JVMs) that are 20 or 30 years old. Now organizations are coming to terms with that to say, Okay, if we want to be in the 21st century, and really help our business take off, we need to modernize our tech stack.”

As to the complexities of microservice management (and how Cortex helps with that), Datta says, “I think part of it also just comes down to being thoughtful about, hey, do I need to build a new microservice for this that already exists in the organization? Secondly, making it clear what the standards are.” He then points to tribal knowledge and how it’s not just “tribal knowledge about what a service does, but tribal knowledge about what service quality means.” He finishes up by saying, “Making it easy for developers to understand, as an organization, these are the things we care about, these are the ways we have defined as the right way to build microservices. I think making that very explicit and codified is supercritical to doing microservices the right way.”

Video Summary was written by Jack Wallen