Top skills IT Ops teams need to manage microservices

Operations teams working as part of DevOps can face significant challenges as they learn to maintain and monitor microservices, especially in a continuous integration and deployment environment.

There are advantages for enterprises adopting microservices over monolithic applications: increased software delivery speeds, improved resiliency, and faster and more effective performance troubleshooting.  

But as DevOps shops make the leap from monoliths to microservices, IT Ops teams need numerous skills to maintain, deploy, and monitor all the services that make up a microservices architecture. Here's a rundown.

Gartner Market Guide for AIOps Platforms

Learn software development

The top challenge for IT Ops teams dealing with microservices is to provide the environments that developers need in a timely manner, said Jay Lyman, principal analyst at 451 Research.

"A lot of this is going to come down to having a little bit more insight or maybe even experience in software development. So I think having software development skills applies."
Jay Lyman

Chris Gardner, senior analyst at Forrester Research, agreed. IT Ops teams are now essentially developers; it's all just code, Gardner said. IT Ops' roles are shifting to modeling infrastructure as code, using APIs all the way down to bare metal, and checking code into repositories, he said.

"This does not mean racking and stacking servers and watching blinking lights, as that was long ago conceded to automation."
Chris Gardner

Specifically, IT Ops professionals should learn basic YAML and/or JSON, understand Kubernetes, and familiarize themselves with distributed tracing tools, he said.

Operations pros should, at minimum, learn a domain-specific language for configuration management. Or if their employers are more advanced in the use of microservices, teams should learn how to create workflows in a continuous delivery/release automation platform.

Understand the development process

At the very least, when it comes to microservices architecture, IT Ops teams need a better understanding of the development process and how things work prior to production, said Ernest Mueller, director of engineering operations at IT security firm AlienVault.

To get microservices-based apps deployed and working in production quickly and keep them up, the approach needs to be comprehensive, he said.

"It's incumbent to understand the whole build pipeline and a little bit about where they're coming from. It can’t be segregated off as much as it was in the old days." 
Ernest Mueller

He also believes that Kubernetes is key. Even if IT Ops people working with microservices aren't using Kubernetes, they should at least be aware of it, and its ecosystem, because it's emerging as a standard, he said.  Netflix uses Kubernetes, as does Google, Oracle, and Amazon. "It's kind of become the microservices platform of choice," Mueller said.

Scale with the infrastructure

With the move from a monolithic to a microservices architecture, IT Ops teams have to be able to scale with the infrastructure, said Amit Kaul, vice president of emerging tech at ThoughtWorks, a technology consultancy.

"You think of something like performance monitoring, and there are tools like Prometheus, there are tools like Datadog that sync well with Kubernetes and container management in general that allow you to do performance monitoring up and down the entire stack," Kaul said.

It's also important for IT Ops to be able to understand multiple languages and tooling, as well as Docker, he said. "You're seeing a lot more cross-language microsystems and cross-language communication, which obviously means IT Ops needs to be able to monitor and debug across multiple languages in a number of different cases," he said.

There's less hard-coupling in a microsystems-based architecture. This means more engineering teams are mapping the right task to the right language, rather than having everybody standardized around Java or .NET, for example, Kaul added.

IT Ops teams are also going to need to help dev teams instrument the microservices for observability at the code level.

"They may need to learn more about how these services can be built with observability built in at the lowest level of the code, and that's where the OpenTracing standards come in. That’s going to be a long-cycle journey—OpenTracing probably won't reach maturity for another 10 years—but the first glimmers are there today, and the standards are out there."
—Gardner

Get in touch with your softer side

Another important ability IT Ops pros need, although more organizational than a skill, is to engage with application developers more directly in integrated teams, according to Gardner.

The idea that there are separate dev and ops teams is quickly disintegrating in modern organizations. In the process, operations pros are learning from developers about agile methodology, languages, and development tools, Gardner said. In exchange, developers learn about standardization, dependability, and security from operations team members.

The end result is a more comprehensive and proficient holistic team.

Charles Betz, principal analyst at Forrester, added that microservices is an architectural style that is closely associated with integrated product teams.

"It comes out of agile, it comes out of DevOps, and the idea is that the team at the end of the day has all the resources and capabilities and authority to actually manage the microservice as they see fit. So trying to take a microservices style into a traditional IT organization with a large, centralized operational team may be a fundamental mismatch—but the jury is still out on that."
Charles Betz

An IT Ops team is also going to need to identify and make the case for any standards required. Consequently, they're going to need some soft skills. They're going to need to persuade microservices teams that architectural consistency is important for operations, and they will need to support people following those standards on a self-service basis, according to Forrester's Gardner.

Beef up your security skills

"Because all of these systems are in a context of increasing social criticality, we have to fix incidents on a more urgent basis," Gardner said. "So IT Ops teams will also need incident response skills. They will need to have the skills to assemble a team and do quick forensics, quick assessment, quick deep-dive analysis on what a likely problem is."

The role of security definitely comes into play for IT Ops teams, agreed Neil Gehani, director of product at Weaveworks.

"They need to understand that the world of microservices has different sorts of skill sets for security. Because now you're not talking about a few large applications, you're talking about a very distributed application with lots of small running pieces. And even though the threat vector may be less, the complexity is definitely more in terms of your overall applications."
Neil Gehani

That means IT Ops teams are going to have to work with both developers and security teams, added 451 Research's Lyman.

And ThoughtWorks' Kaul agreed that cross-team incident resolution and collaboration are critical. "Before, when you had these monolith architectures, you could have one or two people or one team doing incident resolution," he said. Now with all of the microsystems, there are different teams that actually own these different services, products and productions, he added.

When something actually goes wrong, you're now debugging through multiple layers of microsystems and across multiple layers of microsystems, and different teams own those microsystems, Kaul said. "So you need this kind of cross-team collaboration much more than you did in the past."

Don't neglect the basics

In addition to managing the net new cloud-native applications, IT Ops teams must ensure that they have the skills to support both existing and traditional workloads and processes, said 451 Research's Lyman.

"They might need to be familiar with 12-factor application development, which is a type of cloud-native application, and traditional waterfall or ITIL. Those are not going away overnight, so they're still important."
—Jay Lyman

How does you IT Ops team manage microservices? Share your own best practices in the comments below.

[ Upcoming Webinar (Oct. 23): Simplify Discovery and Change Management for Cloud and Container Environments ]