Why automation is critical in devops and digital operations

Presented by PagerDuty

Digital services have cemented themselves as the backbone of every business. But as applications, websites and digital operations have become the main revenue drivers of enterprises, the strain on the teams that maintain these systems has exploded. Last year 83% of developers reported that they’re suffering from burnout.

“The pressure to design great digital services and keep them up and running all the time is greater than ever,” says Michael Cucchi, VP of product and marketing at PagerDuty. “In modern businesses, that pressure falls on software developers. It’s new skills, new time commitment and working with new parts of the business they didn’t have to work with before, and that causes headwinds and strain as they migrate into that environment.”

Constant interruption is the most impactful factor in developer burnout. Studies have found that interruptions cost the U.S. economy $588 billion a year, steal three to five hours a day at work, destroy concentration and harm morale. Every time a developer is alerted to an issue that turns out to be a false alarm, or something that’s outside their scope, productivity drops and frustration grows. In PagerDuty’s State of Digital Operations report, more than 60% of those surveyed reported responding to off-hours alerts once a week or more, and can add up to 12 extra work weeks every year. This means they are more likely to get burned out quickly and leave.

“The most skilled developers, the most prized people that build the best code the fastest, are actually the ones that bear the brunt of the operations load,” Cucchi says. “They are trying to build better services so they are more resilient, yet they’re constantly distracted to go potentially troubleshoot a false alarm. They’re losing sleep and losing weekends. They’re the people you need at the company the most and ironically, the ones that are most prone to leave.”

Cucchi adds that the key to addressing burnout is democratizing the skills of those developers — if only one developer has all the answers, scaling becomes impossible. That scaling requires digital transformation: building automated intelligence in your environment so that almost anyone can securely fire off the response that solves the problem without having to wake up a specific individual because they’re the best developer in the organization. Process automation like this can also have massive positive impacts across the whole organization.

How automation transforms devops

At the top of the list of benefits, automation empowers developers. It gives them self-service access to all they need to get coding and deploy very quickly, as well as rollout and scale. But the highest-impact area for automation is the rapid and compounding ROI it can enable when it dramatically reduces the time and effort required to address the critical issues that can damage a company — if an ecommerce website goes down for example, millions in sales can be lost.

“Automation is big in terms of empowering developers to get the services and the skills and the tools and all the resources they need, but then there’s this other side of things, which is acting as a shock absorber,” Cucchi says. “It’s distributing developer knowledge and control in a secure way, so you can automate a problem response in real time, or give your lowest level responder the access they need to securely run administrator behaviors and solve problems — without having to wake up the senior developer three time zones away.”

An IDC report found that PagerDuty’s own customers have realized $3.48 million in average annual cost savings, a 795% three-year ROI, and just two months to full payback on the investment in automation.

And the benefits extend across the entire organization. With the company’s acquisition of Catalytic, PagerDuty Process Automation is for more than just developers. Sales teams, customer service, HR teams, legal teams and other business units will also be able to create flexible workflows to tackle repetitive work and free up time. The first of these was recently announced, Incident Workflows, which     are designed to dynamically respond to problems by escalating issues to the right team and automatically kicking off response actions like creating a Slack channel, sending out an email to the organization or running automated diagnostics.

When the troops across the organization are given the power to solve problems and take issues off the developers’ plate, developers get to code for 50% more time. And noise disappears, as machine learning and automation reduces sometimes over 90% of false alarms, without having to interrupt a human to manage the issues.

There’s also a major reduction in time it takes to solve a problem — PagerDuty customers have seen a 50% to 80% drop in the duration of incidents. That’s because an AI-powered automated platform is intelligent, learning as it attends to issues that crop up, and coming armed to each problem with the information required to resolve it successfully.

With machine learning, an automated platform can ingest change events from code deployment     pipelines and correlate the change and identify who made it. You know who to wake up, and with automation, you can roll back to the last state automatically, without a human and in less than a second.

Starting the automated devops journey

The easiest way to get started is actually in the lowest-risk places, Cucchi says, such as automation for batch processes, staging or provisioning, standing up new environments, running diagnostics, discovery tasks and so on.

“The key is getting comfortable with it and getting a unified way to manage and deploy it to your entire enterprise,” he says. “Then you start to see the ROI take off. It’s not just standing up a cluster in a cloud. It’s running a whole business process for your enterprise. It’s offloading tens and hundreds of hours of skilled labor when you move into that direction.”

Start by implementing a process automation platform that is secure and scalable, easy to deploy, use and maintain. From there, move it into higher return places, where your company’s really hurting and minutes matter.

When implementing a process automation platform, you also need to understand the personas across your company — you don’t want to give the keys to the castle to everyone, he adds. It should be secure and highly authenticated, with tiered access privileges and capabilities and able to automate across any environment you choose, cloud or otherwise.    

“You want to set up your customer service team with routines that keep them informed and solve their problems and give them proactive interactions with the customer, but obviously you don’t want them to be able to take down the primary database,” he says.

Each persona gets an individual, tailored interface, a real-time status dashboard, and access to automation routines with pulldown menus. A customer service agent would have access to Salesforce; a tech troubleshooter would have more granular access to the IT stack.

“It’s building these tailored experiences, and then delivering them in a way that’s consumable to that user,” he says. “The business should design automation in a way that’s not intimidating, that’s just straightforward and easy, empowering employees as you introduce them to it.”

Many companies have already invested in siloed, bespoke automation in their environment, whether it’s virtual or cloud infrastructure, code process rollouts, the CI/CD pipeline and so on, all in the low-risk, but just-moderate return category. But the massive transformations happen when you go from just using automation for these kinds of mundane, offline tasks, to moving it into real time — and that can help keep developers at their company.

“It’s critical to stabilize how the ecosystem is operating for your developer, and then accelerate your innovation rate,” Cucchi says. “If you try to drive developers into advanced devops methodologies without giving them these technologies, tools and communication methods and offload methods, they break. That’s why it’s been a decade and a half of every company in the world trying to master devops as an approach. But automation, machine learning and AIops is delivering here and making it a reality today.”

Source: Read Full Article