How do you make your IT more iterative and responsive? How do you increase collaboration between support and development?
High-performing IT orgs are 2x more likely to exceed productivity goals.
Bringing DevOps principles into your IT support and engineering teams dramatically improves service quality, team morale, problem-solving and business productivity.
In this webinar, ITSM expert John Custy shows you how to apply DevOps principles to your IT org. This presentation is for anyone involved in the support and development of IT systems and services. The keys to higher-performing services are so simple, they might surprise you.
In this webinar, you'll learn:
- Best practices for organizing your IT and internal development projects.
- How to bridge development and IT.
- Key metrics for tracking DevOps success.
Watch the webinar now
Your questions, answered
Our Q&A team chose the top 7 questions from the audience to answer in this article. Enjoy!
1. How would DevOps be run in an organization with a head of Engineering and a head of IT?
The first challenge is getting the head of engineering and the head of IT to agree on a common set of goals. For instance, engineering should have some reliability goals. The key is to identify responsibilities of each group in terms of the whole cycle.
2. We are moving our IT applications to the cloud - would we still use the ITIL framework or should we shift to DevOps?
You would continue to use the ITIL framework. However, the ITIL framework doesn't really address a lot of software development activities. It tends to address the design, but doesn't get into development. Here, you wolud bring the concepts of lean and agile to ensure all development projects are done in a timely manner. It's ideal that you bring the two together. One question many people have is, "Is ITIL lean or agile? Or is it just very bureaucratic?" By design, ITIL is agile. The problem is how it's practiced, which might make it bureaucratic. But, that's not the way the ITIL framework was designed to be implemented.
3. Should developers have access to production servers?
Typically no, they shouldn't. However, this question brings us back to these three practices. Continuous integration, which is about developers committing their code to a shared respository. Continuous delivery, which is about ongoing testing and should be automated. And lastly, continuous deployment, which is about getting things out to production. I wouldn't have developers have direct access to it because we need to make sure we have the right controls in place. We need the gates, but those gates should be automated with criteria known in advance so we can move through them. The gates should not block or cause delays, but they are important checkpoints.
4. What's the difference between continuous delivery and continuous deployment?
Continuous delivery is making sure that the software is always in a releasable state. Continuous deployment is about making sure that all of these changes are actually deployed. Think of continous delivery as "I'm ready to deploy" and continuous deployment as "I am actually deploying these changes to the infrastructure." This question deserves a webinar on its own, but that's our simplest answer.
5. At what size does an organization need to implement DevOps?
The core DevOps principles apply to the smallest team up to larger organizations, like Amazon, Netflix, Google with thousands of developers. The differences might be in how much automation you have or how many different sets of tools you have. But culture, flow, feedback, communication, collaboration, continual experimentation - those apply to one person, three people, five people.. all the way up to 20,000 people. How formal the controls are will obviously be less with a smaller organization.
6. What tools do you recommend for a DevOps-oriented IT org?
There are tons of vendors out there and everyday, they're changing. I like looking at tools for their purpose. Look at tools for automated testing, understanding changes, monitoring, dashboards, production logs, measuring process efficiency, compliance, incident tracking, knowledge. If you're looking for tools, ask yourself these questions: Does it work in our environment? Does it meet our needs? Do the tools work together as a comprehensive, cohesive toolset? Atlassian's products integrate together and can be used in a DevOps environment. Use JIRA Service Desk for incident and change management, JIRA Software for software development, Bitbucket for your code repository and HipChat for team communication. The great thing about HipChat is its integrations with monitoring applications and bots that wil even run automated tests, deployments and environment setup.
7. What should IT support be automating?
There are three areas where support can automate. One is workflow, and moving tickets through the process faster. The second area is knowledge. When an incident comes in, the service management tool should be searching for knowledge and identifying related articles. The third area is escalations. Automation can help IT move beyond the traditional escalation mindset to something more intelligent. Often, there are issues that come in that only 1 or 2 people can solve. Instead of escalating right to those experts, traditional escalation paths make you go through a predefined set of steps. With smarter tools, you should automatically find and escalate to the right subject matter experts.Try JIRA Service Desk for free