We had our product, the business, the software and hardware architecture ready. We had tested some elements on our laptops. We had tested individual servers. We were ready. The only missing thing was an approval stamp from the SRE team.

At the time, the buzz was a book from Google, The SRE one. It was taken for absolute knowledge. Scaleway’s SRE had been created to imitate Google’s. They were supposed to handle every product’s production environment. To do that, they required that they approve every product’s design.

Context is important there. Different teams had been created to launch new products. The responsibility of handling the existing ones had fallen upon us. Six months in, we still had to maintain the existing products, SRE had not yet taken their load.

The implicit contract as we understood it was that each team was almost like a startup within Scaleway. For me, for Storage, we were doing exactly that. Product & Business. We made our choices. We justified each choice, software and hardware.

With that context in mind, you can imagine our reaction when we were told that they would not stamp our design. They required that we change a large portion to be more “SRE compatible”.

A conflict erupted. Which is pretty normal. Conflicts happen. I wish at the time I knew better how to handle this, but that is how you learn. I just couldn’t believe that the team created to handle production, which had up to then, refused to take the load, was giving us lessons on what worked or not.

Just because a big tech company publishes a book doesn’t mean you use it as absolute truth. Google in 2017 had 80 000 employees. Scaleway was growing fast, we had doubled our head count to 280 by that time. The last time Google was at 280, was more than fifteen years back, in 2001.

Even by taking a conservative count at 20% engineers, that was still 1 600 engineers, far more than the total combined workforce of Scaleway. Even worse, they had compounded engineering tools, knowledge and expertise for twenty years. Many things Google published were not relevant to us. It could not be applied in our context. It was like applying a car factory solution to a car workshop. Or using EUV to etch 100 micrometers chips.

Yet, I saw people blindly applying concepts. As an individual contributor and as a leader, choices you make depend on context. Make the best of it. Don’t just copy.

The SRE team did not budge. We didn’t either. This wasn’t moving in either direction. No one was giving anything. Eventually we got approval from the executive level and got approval on the condition that we would maintain the production ourselves. We were already doing that on the existing products.

The situation unlocked. The conflict was now resolved. We were ready for the next level. We needed the foundation.

(1) Photos by the launch event photographer - if you know who it was please tell me

If you have missed it, you can read the previous episode here​

​

To pair with :

• Waulking Song - TEED, Lone remix
• Skunk Works: A Personal Memoir of My Years at Lockheed by Ben R. Rich, Leo Janos