SRE at vpTech: Our Solutions for a Seamless Workflow

vpTech has successfully established an open source-like culture for the SRE teams, optimizing the engineers’ workflow.

“Is being an SRE stressful?” is the first Google suggestion that pops up when you ask about the challenges of the job. Indeed, the job description comes with serious responsibilities and juggling numerous tasks.

Could it be that the way to make the full SRE experience smooth and efficient is to bridge the gaps in communication among the teams?

It turns out there is a way to make an SRE experience smooth and efficient by bridging the gaps in communication among the teams. This is not an article about the usual SRE challenges — this is a detailed explanation of how vpTech has worked hard for the last two years to make the SRE experience at vpTech well-ordered and varied.

Let’s see the 50 SRE specialists at vpTech work today, given the challenges of the job.

Unified catalog — a breakthrough for communication and productivity

The SRE department at vpTech consists of 6 large teams, and without proper communication, they could end up losing time and resources on similar tasks. Often, one team can already have experience and code for the job another team is handling.

“We wanted to have time and energy properly used, as well as provide our SRE with the fun part of sharing their knowledge and proficiency with the team. It has been a massive amount of work, but we managed to assemble a unified catalog of SRE resources for vpTech”, says Antoine Millet (Principal SRE), who launched this challenge in early 2020. Two years later, each team member has access to a massive codebase and numerous resources with an intelligent search interface where they can get all the information on a project.

The unified catalog is an impressive invention and a massive time-saver for engineers. It also boosts communication among the members and clarifies the work processes. During the build phase, one may get an impression of being isolated, while the catalog offers balance and good energy.

Theory aside, here is what the current vpTech catalog contains:

• 1 Ansible collection
• 200 Ansibles roles (not playbooks)
• 80 GitLab CI/CD templates
• 30 Debian packages (mostly open-source prometheus exporters)
• 100 Container images
• 65 Helm charts
• 6 Prometheus exporters
• 11 Python packages
• 20 Terraform modules

Workload distribution — finding the perfect balance between run and build

Another challenge, according to the SR engineers, is having to take care of recurring and unpredictable tasks manually, and this is the least exciting part of the job. One SRE team at vpTech works with 20 to 35 product teams, who, in their turn, can reach out with the following requests:

• Questions
• Tickets
• Tasks
• Minor fixes
• Security fixes
• Asking for access
• Monitoring updates

Site Reliability Engineers split their time among building platforms, operations, and maintenance. Meanwhile, any task that has not been part of the roadmap due to its unpredictability can show up on the to-do list.

vpTech’s goal is to ensure that the workload is distributed smartly, where each engineer equally contributes to recurring or new tedious requests and has the freedom to organize the rest of the workload.

For example, for a team of 5 people, one person will take care of the “RUN” tasks one day per week. Like that, not a single request is left unattended, and it also removes the pressure of having regular monotonous work on the agenda. vpTech values smart workload distribution and optimization since they have been shown to contribute the most to performance and work energy.

SRE is challenging but can also be exciting, with many opportunities to grow with the company. With the unified catalog and transparent communication, the engineers at vpTech can streamline repeating tasks and have the flexibility to apply their skills at what they do best.

If SRE per se benefits the performance and speed of the product, shouldn’t we start by offering our specialists tools and work conditions to enhance their performance and speed in the first place?

This was the idea back in 2020 for vpTech, and this is exactly what the company has accomplished today, with many more improvements to come.