#8 Site Reliability Engineering - 1st meetup of 2021 🎉

Bonjour à tous et à toutes,

Nous sommes ravis de vous inviter au 8ème meetup SRE Paris et premier de l'année 2021 🎉

Cet édition de janvier aura lieu complément en ligne avec un programme bien complet et intéressant, avec des retours d'expérience sur Networking et Load Balancing sur GKE, sur le développement des Operators Kube et la présentation d'un projet lié à l'Infrastructure et energy consumption.

Voici le programme de la soirée :

• 19h00-19h10 : Présentation du programme et des news du groupe
• 19h10-19h30 : Building Custom NAT for GKE par Brice Santus (@azraeht) / Wenceslas Danguy Des Deserts
• 19h30-20h00 : Hello, Operator? Give Me a Resilient Data Processing Pipeline par Arthur Busser (@arthurbusser)
• 20h00-20h10 : Scaphandre: understand and shrink the energy consumption of your infrastructure par Benoit Petit (@bpetit_)
• 20h10-20h20 : Live demo of driftctl par Stephane Jourdan (@sjourdan)
• 20h20-20h30 : Questions et du feedback pour le prochain meetup

A bientôt et bonnes fêtes de fin d'année !

# Détail des présentations

## Building Custom NAT for GKE

expected time : 20m
author name: Brice Santus (@azraeht) / Wenceslas Danguy Des Deserts
author bio: SREs @ Cybelangel
language: FR/EN

GCP offers a powerful but expansive Cloud NAT product which is fully managed. After several months using it, it appears its cost was too high for this feature simple feature. We decided to build our own NAT gateway on top of our GKE cluster.
This the journey of our setup, what are the reasons that led us to build this solution. How we reduce our inbound/outbound network cost of 95% using few architecture tweak and less than 50 line of golang.
In this story we will talk about network problems, dynamic loadbalancing, golang and IAC.

## Hello, Operator? Give Me a Resilient Data Processing Pipeline

expected time: 30 minutes
author name: Arthur Busser
author bio: Arthur is an SRE @ Padok and a Certified Kubernetes Administrator.
language: FR/EN

While building a platform for on-demand data processing pipelines,
Padok wrote a Kubernetes operator to answer specific scheduling needs. This allowed their SREs to maximize resilience while minimizing resource waste. In this talk, Arthur will share his experience in developing this operator, including what problem it solves, what open source tools in was built with, and why it was definitely worth it.

## Scaphandre: understand and shrink the energy consumption of your infrastructure

expected time : 10m
author name: Benoit Petit
author bio: Founder @ Hubblo
language: FR/EN

We are facing a climate crisis due to... ourselves (I mean mankind, not only SREs). One vector of this crisis is greenhouse gazes emissions.
One factor of those greenhouse gazes emissions is electricity consumption (electricity doesn't appear magically right ?) and our industry consumes quite a lot of electricity.
Let see how we could tamper that, using the best collaboration method we have: open source and open collaboration.
I'll present the scaphandre https://github.com/hubblo-org/scaphandre project and it's perspectives.

## Live demo of driftctl

expected time : 10 mins
language: FR/EN
author name: Stephane Jourdan (@sjourdan)
author bio: CTO and entrepreneur

Infrastructure as code is awesome, but there are too many moving parts: codebase, state file, actual cloud state. Things tend to drift. Drift can have multiple causes: from developers creating or updating infrastructure through the web console without telling anyone, to uncontrolled updates on the cloud provider side. Handling infrastructure drift vs the codebase can be challenging. We will perform a live demo based on Terraform and AWS of driftctl, a free and open-source CLI that tracks, analyzes, prioritizes, and warns of infrastructure drift.