Five challenges for enterprises to build a successful Devops model


[editor’s note] the author of this article is Kevin Goldberg. It mainly introduces the five challenges and problems that enterprises need to deal with in order to successfully deploy Devops mode. This paper is a compilation and presentation of the domestic itom management platform oneapm.

It’s not easy to give Devops a concise, accurate and appropriate definition. However, I have seen a sentence in the past that seems to better explain what Devops are – “Devops is a culture, movement or practice that emphasizes communication and collaboration between software developers and other IT professionals to jointly facilitate the automation of software delivery processes and infrastructure changes.”

Now that you understand what Devops are and what is not, you have to think about why your enterprise should use Devops? In addition to breaking the independent working model between two decentralized departments, what other significant changes and achievements can you expect? According to the latest Devops status report of puppet, they found that “efficient IT organizations experience 60 times less failures and 168 times faster failure recovery than inefficient it organizations. Moreover, their deployment time is 200 times shorter but their deployment frequency is 30 times higher. ” Simply put, a successful Devops organization can achieve faster product release frequency and fewer problems in a shorter time.

To achieve a more collaborative Devops transformation, it is important to understand what potential challenges the company may face before you start.

1. Don’t start from top to bottom

This is one of the most common problems many enterprises encounter when trying to implement Devops internally. In the actual process, it is impossible for any executive to directly copy a successful Devops model, and start to implement it in the whole company at a command. Because it can’t be carried out from top to bottom.

A successful Devops campaign should be initiated by “grassroots” and then evolve into a company wide cultural change. The perfect situation is that several development and operation and maintenance personnel work together to deal with some small events and establish a good understanding in the process of cooperation. Then the new team atmosphere gradually heats up, naturally infiltrates into both departments, and finally forms a step-by-step feedback process and work rhythm.

2. Start with the right project

Many large and medium-sized enterprises have employees and offices all over the world. There is no doubt that this poses a huge challenge to collaboration. Because no matter it’s language barrier or time difference, or just because of the decentralization and independence of each department, the larger the company is, the more difficult it is to form a perfect Devops model. That’s why it’s important to use the right tools to facilitate collaboration (more on that later). It is also important to choose the right project when you start deploying Devops.

Through the use of various tools, if your Devops team moves forward step by step, although the progress is slow, but substantive results have been achieved, then team members around the world will pay more attention to this new work mode and are eager to join it. The quality of the products completed by the operation and maintenance team will be better, and all team members will be proud of it. Especially after seeing the results of their work, each member will have a more sense of responsibility, and the next work will be harder.

3. Innovation of strategy, cloud and tools

“Legacy”, in this context, refers to the strategic and instrumental challenges that stand in the way of the DevOps movement. The first and most important is the legacy of the team atmosphere and culture. We often hear management say, “we always do that.”. Yes, that’s why at the beginning of the article, I emphasized that we should promote Devops movement from the technical level of the company rather than the management level. Because the following people need and want to develop to Devops, once they have achieved results, there will not be too many objections.

The second legacy is about existing tools. If your team is still using some antiquated tools that are slow to release and not agile, then you can start to discuss the migration of tools. More and more enterprises have invested a lot of money and trust in cloud and integrated products. When deciding which tools to buy, be sure to consider whether they match the application’s lifecycle and whether they work best with other tools.

4. Quantitative results and impact

Another challenge we often hear about is that many successful Devops teams in an enterprise are hard to get management support because their results are hard to quantify. Although puppet’s survey report shows that Devops has achieved great results, the company’s management may not think so. As a leader in the Devops team, you need to consider the following:

  • How often does your team release updates?

  • How many votes did you get after the update was released?

  • How often is an app released on average?

  • What are the benefits and / or brand impact of each release?

These questions help you assess your current release process and performance and identify where you can best improve. Once you have identified these issues, you can take this KPI data that is most suitable for the company’s situation to talk with the management team about your results.

5. Select a toolset for the entire lifecycle

A few weeks ago, I wrote an article about how to choose the best tools and how to map them to the application lifecycle. In fact, this strategy can’t be overemphasized, because we often see that some teams choose tools just because they look cool, or some teams want them, but they don’t use them for various reasons.

If you are interested in learning more about Devops toolkit, please click here to read our free e-book.

This paper is compiled by oneapm engineers. Oneapm cloudinsight integrates monitoring, management, computing, collaboration and visualization to help all IT companies, reduce the human and time cost investment in system monitoring, and make the operation and maintenance work more efficient and simple. To read more technical articles, please visit oneapm’s official technology blog.

This article is from oneapm official blog

Original address: