Customer stories Veepee
+

How Veepee accelerated deployment from 4 days to 4 minutes

Native CI integration
Improved workflow
Accelerated deployment
Industry Retail
Employees 6,500
Location La Plaine Saint-Denis, France

Want to see what GitLab Ultimate can do for your team?

Start your free trial

Veepee completely overhauled its central repository and increased deployment and security with GitLab.

GitLab Enterprise helped transform Veepee’s IT production with enhanced continuous integration (CI), automated testing, and expedited continuous delivery (CD) for improved process workflows.

Veepee’s availability rates went to 99.98% due to the faster feedback cycle and accelerated triggers.
- Antoine Millet, Head of IT Operations, Veepee

European e-commerce company

Veepee, formerly known as vente-privee.com, was founded in 2001 as an e-commerce company that specializes in flash online sales. Veepee creates daily sales events in partnership with leading brands, on a limited-time basis with specific products, sold at heavily discounted prices for industrial or promotional reasons.

Based throughout Europe, Veepee represents the successful transformation of a long-established activity — clearance sales — through a revolutionary e-commerce model that has since been copied worldwide. vpTech, the Veepee Tech community, supports the business and the growth. The team handles more than 70 products using around 50 technos. As a European company spread in six different countries, they advocate feature teams that are flexible, autonomous, and multicultural.

Spreading the inner-source philosophy to improve overall delivery process

The year 2016 was a turning point for Veepee that included expanding the company throughout Europe. To get ahead of any upcoming issues, the vpTech team came with a strong promise: Align all Veepee legacies and adopt a modern and agile information system. To do so, the team grew up from 300 to 750 people.

At the beginning, vpTech faced several challenges with its initial release workflow. Each team worked within their own toolset, which made collaboration and contribution difficult. Teams were frustrated that they couldn’t easily access other’s code and had no way to successfully collaborate with others.

Also, with the previous workflow, deploying to production was a slow and tedious process. “It was not an ideal situation at all, as it was nearly impossible to release something fast — only emergency cases were free from the whole process,” explained Antoine Millet, Head of IT Operations at Veepee.

vpTech was looking for a way to modernize its engineering infrastructure and enforce inner-source. At the same time, top management wanted to keep code and instances in house and stay off the cloud for the time being. CI integration was a key requirement for the development teams.

GitLab and its complete toolsets

vpTech chose to fully reboot its tooling and workflow to modernize engineering tools within the company in order to allow teams to work in a singular environment with common tool sets and shared methodologies. vpTech researched and ended up adopting GitLab. “Three years ago, the biggest difference between GitLab and GitHub was the native CI integration. That's something we really needed at Veepee,” Millet said.

The team adopted GitLab rather quickly. The initial goal was to start with 300 licenses over the span of three years. However, developers gravitated toward the platform and demanded more within the first six months of the adoption period. Within a year, vpTech had over 1,000 licenses.

vpTech set out with new technology and deployment initiatives. The first initiative was to get developers onboarded to GitLab and completely off of the previous workflow. The second was to explain and spread the inner-source philosophy: If you need something, contribute. The third was to ensure that if a deployment in production happens, testing occurs in the GitLab CI pipeline for visibility.

A new workflow, a changed culture

Engineering and product team workflow has improved significantly. If the team wants to change a form on the website, or introduce a new A/B test, they simply release code. If the test passes, they can deploy in production. “Before GitLab, it could take up to four days for a production deployment because we had to involve a lot of people. Right now, it's a few minutes without any human in the middle. It's fully automatic,” Millet said.

The technology and the deployment process with GitLab CI belong to the product team, so they are empowered with choices for how they deploy. “We produced the last component of the CI, the deploy part. We created tools for engineering which will automatically execute blue and green deployment for them to try everything through canary deployment. If it's validated, it’s deployed in production. That's the front workflow,” Millet added.

The site reliability engineering (SRE) team produced a lot of CI templates, which created a center of knowledge for the teams. If there is a new developer in the company, templates can be easily accessed in order to get started. “We have sets of linter for all languages pre-templated by the SRE team. Basically, you just choose that first component. In the middle, for the test, you put your own code because you know what you did, so the test part is for your team,” according to Millet. The deployment is templated for either Nomad and Kubernetes to get in production.

In the last three years since adopting GitLab, Millet has never received a single complaint about the platform. GitLab significantly changed the culture of a 16-year-old company by improving transparency, workflow, and developers’ time. “It took almost a year for people to adopt GitLab as it was a change of habit, change of the way to work. Now it's working really well. At this point, we have developed a tool named DevHub on top of GitLab to make statistics and give a full view for the top management,” Millet said.

All information and persons involved in case study are accurate at the time of publication.

Ready to get started?

See what your team can do with the most comprehensive
AI-powered DevSecOps platform.