Over a million developers have joined DZone.
{{announcement.body}}
{{announcement.title}}

DevOps Starts With Leadership and Culture

DZone's Guide to

DevOps Starts With Leadership and Culture

DevOps is more than tools — it's about building a culture of collaboration. See how you can foster this as a leader in software development.

· DevOps Zone ·
Free Resource

Read why times series is the fastest growing database category.

DevOps is truly successful when it is more about building a culture of collaboration and continual automation and improvement than it is about expertise in specific toolsets.  No doubt, the tools are important, but DevOps leaders leave a lot of value on the table if the culture does not fit.  It’s up to the DevOps group leaders to build that culture.

One of the first major activities for DevOps leaders is to get buy-in from all the stakeholders and to make DevOps as minimally disruptive as possible.  First, they start showing the value of what DevOps brings to the organization. Then they start looking at the staff levels, and the qualifications of folks involved.  They bring a point of view that good DevOps resources may have strong infrastructure and development skills, but also have to have strong troubleshooting and problem-solving skills. They want people who want to use data as one of the key elements to success so that it’s not just about getting applications to work but to make sure the process gets improved and automated so that the applications work better and require less management.

Designing the Organizational Model

Successful DevOps is not just having DevOps engineers come in and use automation to provision infrastructure. It's connecting the DevOps activities to what the development engineers and operations folks are doing, and building bridges across all functions and groups. Leadership will break down silos, and in breaking down barriers between those specialties, a powerful organizational structure will emerge.

Leaders will need to recruit and empower architects, managers, and engineers and decide when to expand the team, and how to make changes to what people are doing and maybe broaden or narrow their roles.

They will also have to come up with a timetable and process to mature the tools, standardize on tools, and build automation, orchestration, CICD, and containerization workflows.

Building and Measuring the Process

It helps to be an agile organization to fully embrace what DevOps can bring to an organization’s culture.

Having the right technology, whether it's where the resources are hosted, or the right toolchain is key achieving DevOps promises. In addition, the teams need to measure what they are doing in order to improve reliability and throughput. For example, leaders often want to have the right metrics in place to be able to show how many builds got revved last week and out of those builds, how many got successfully deployed to production. Even more importantly, they will want business metrics as well. 

For example, they will want to know how a deployment yielded an increase in a number of orders, depending on the feature. Or how it helped cut back on the number of operational elements that cause unreliable systems to fail. Leadership should look at all the different data points, whether it's systems, development metrics, business metrics, and aggregate them and put them in the KPI. It’s not just about uptime, downtime, what is CPU or storage costs this week or this month. It's all about whether the application as effective as it could be as measured in business terms.

Picking the Right Tech and Tools

Leadership needs to make sure the tools and engineers are supporting what they're trying to do.  They need to encourage the engineers not to go off and do their own thing, but bring any required feedback back to the release engineering function, have them buy and build tools, and collaborate, in order for the organization to have the resources available to satisfy the challenges it is running into. Once the DevOps group has created a toolset and deployed and gotten the adoption they want to the DevOps process, a leader makes sure that an engineer married to the program, and they own it for the entire lifecycle of that product.

DevOps engineers with the right skill sets are not inexpensive. For example, leaders want to give the site reliability engineers and the operation folks all the dashboards they need. They want to articulate an understanding of what's important and what's noise so that they can effectively bring reliability to the product.  DevOps as a culture ties software engineering functions nicely together because it allows the organization to scale better using the self-service tools.  The individual functions can come together and provide a more holistic solution.

Recruiting and Managing People

When building their organizations, successful DevOps leaders often start by picking people who are more collaborative in mindset along with having the engineering skills to use collaboration tools.

DevOps engineers are no longer just an infrastructure engineer that wants to do development or a developer that wants to do infrastructure. Effective leaders are building teams where many effective DevOps people come from a data background with software engineering.

Many leaders believe that the most effective DevOps engineers have an understanding of not just infrastructure, networking and security, and data, but their knowledgebase includes middleware and applications.

Breaking Barriers, Building Bridges

With DevOps, good leaders know how to break down barriers, and bridge engineering with operations. That's because the organization wants the DevOps engineer, or the program itself, to have ownership, not just of the features they're releasing, but how it performs in production.  That makes ownership and accountability a key attribute.

The leader needs to not only communicate with the team but instill that sense of ownership.  If something is constantly going down, it's their responsibility to make their jobs easier, and make engineering's job easier, and address those problems.  This way the feedback loop from product back into the engineering becomes very valuable. It's important, and that's what the leaders should be looking at and setting up for foundationally.

Learn how to get 20x more performance than Elastic by moving to a Time Series database.

Topics:
devops ,leadership ,agile

Opinions expressed by DZone contributors are their own.

{{ parent.title || parent.header.title}}

{{ parent.tldr }}

{{ parent.urlSource.name }}