Approach to GitOps: achievements, mistakes and lessons learned. Targetprocess experience
GitOps has been around for several years already. In fact, it’s not that neoteric, since it’s mostly good old Infrastructure-As-A-Code and Configuration-As-A-Code principles being applied to microservices/cloud-native/k8s world. Pretty simple, isn’t’ it? But, there’s always a “but”.
This talk will cover our experience with building Continuous Delivery via GitOps here, at TargetProcess. We’ve been using this approach in production for some 3 years already, being able to (more or less) successfully manage umpteen microservices in dozens of production environments. Our stack is terraform, docker, k8s, helm, nexus, git(lab) and friends (where “friends” stands for our internal developed set of tools that puts all the pieces together).
We’ll start with a brief introduction to GitOps itself, it’s history and main concepts. Then we’re going to recollect our own experience with GitOps: the background, architectural decisions, technical stack, custom development, pitfalls, wins and failures – everything that we’ve faced during recent years. Finally we will address the global picture of GitOps (well, at least the part of the picture that we suppose to see).
Infrastructure developer, Targetprocess, Inc.