🚀 Next up: ALEKSANDR PLOTNIKOV
Staff Cloud Engineer at inDrive

Aleksandr Plotnikov works with a multi-account AWS estate at inDrive, managed as a Terraform/Terragrunt monorepo of around 14,000 modules.
At that scale, even a simple question becomes a serious engineering challenge:
When a pull request touches ten modules, which one applies first?

For years, the answer was orders.txt — a hand-maintained file defining the global apply sequence.
It worked, but it was slow. Average apply sessions took 22 minutes, with the 90th percentile reaching 49 minutes.
So the team deleted the file.

🗓 On September 25, Aleksandr will share how they replaced it with a computed dependency graph and unlocked safe parallel applies.
"We Deleted the File That Ordered Our Infrastructure — and Applies Got 2x Faster"

The result: 2x faster average applies, up to 2.5x faster in the worst cases, and up to 9x faster for wide infrastructure changes.
More importantly, he’ll show what it takes to make parallel infrastructure changes trustworthy at the scale of 14,000 modules — based on six months of production data at inDrive.

🗓 𝗔𝗪𝗦 𝗖𝗼𝗺𝗺𝘂𝗻𝗶𝘁𝘆 𝗗𝗮𝘆 𝗖𝗲𝗻𝘁𝗿𝗮𝗹 𝗔𝘀𝗶𝗮, 𝟮𝟱 𝗦𝗲𝗽𝘁𝗲𝗺𝗯𝗲𝗿, 𝗧𝗮𝘀𝗵𝗸𝗲𝗻𝘁.

⏲ 09:00–18:00 at CAEX.

🆔 Register for FREE

Visit awsug.uz for the full program.

🔗 Connect with Aleksandr Plotnikov:
📱 LinkedIn