A new report from software analytics firm OverOps found that while DevOps adoption continues to grow, it’s actually leading to “reliability chaos.”
Standing in sharp contrast to the goal most organizations have when adopting DevOps, OverOps’ report details an environment in turmoil: Developers are still spending lots of time troubleshooting, accountability is lacking, and quicker deployment times are leading to production errors that could have been avoided.
All this doom and gloom can make it seem as if the report focuses on the failure of DevOps to revolutionize the enterprise, but that isn’t necessarily the case. As with the implementation of any organizational structure, DevOps has to be done correctly; when it isn’t, there’s sure to be problems.
Luckily for DevOps organizations sympathetic to the findings of this report, there are solutions.
What’s happening in DevOps?
If the report is to be believed, asking the average developer or IT operations professional if they’re using DevOps would yield a result that makes it look like an unparalleled success: Some 75.9% have partially adopted, or are considering adopting, DevOps;and 17.8% have fully adopted it, making the total percentage of companies in the DevOps realm 93.7%.
SEE: IT leader’s guide to making DevOps work (Tech Pro Research)
Along with increased adoption of DevOps comes increased harmony between developers and operations: 73% of those surveyed believe application and service quality is a shared team responsibility, and 66.9% say the whole team is to blame when something goes wrong.
Decreased finger pointing and an improved sense of team are great results of implementing DevOps, butfrom here, things start to look a bit grim.
The heart of DevOps darkness
With DevOps a nearly universal concept in the modern enterprise, it stands to reason that there are going to be issues. If so, the numbers in OverOps’ report indicate there’s more than just a margin of implementation error at work: Something is going wrong in a lot of DevOps organizations.
Take automation, for example: DevOps is designed for faster release schedules, which means that automated tools are used to catch an increasing percentage of software errors. Despite increased use of automation, 76.6% of respondents said they’re still using manual processes, and a shocking 52.2% rely on customers to tell them about errors.
All that manual troubleshooting takes time, with 20% of respondents saying they spend one full workday a week fixing bugs, and another 42% spend between one half and one full day.
Think back to the shared responsibility that developers and operations feel under DevOps, and you can start to see where OverOps’ report is going: The big problem in DevOps is confusion.
The biggest concerns that DevOps professionals have over their processes all come back to one thing: confusion.
DevOps is supposed to make the software lifecycle more flexible, products more reliable, and teams more collaborative; but without clear marching orders, it’s impossible for a team to do that.
OverOps calls one of the biggest problems facing DevOps the “multiple owners = no owners” syndrome, which is something IT managers and lead developers need to nip in the bud before it can take root.
DevOps is a process, but it’s also a philosophy that has to change the entire way work gets done. Success means having resources available, roles assigned, and structures clearly outlined before even considering implementation.
If you see reflections of your organization in this report it’s not too late, but don’t be afraid to tear things down and start fresh: DevOps success starts at the foundation.
The big takeaways for tech leaders:
- Nearly all companies have, or are considering, implementing DevOps, but many are experiencing problems with their implementations that are creating organizational chaos. — OverOps, 2018
- DevOps professionals say confusion is the biggest problem they face. A successful DevOps implementation has to have clear standards, roles, and responsibilities assigned to all teams before implementation begins. — OverOps, 2018