Consul: Up and Running

Preface

The sheer volume of software required by today’s world has necessitated an evolution in how we structure our engineering organizations. We’ve learned that smaller, independent teams work better than larger, highly coupled ones. Since Conway’s law—that companies will produce systems to match their organizational structure—is inevitable, this evolution has precipitated the rise of microservices: smaller, independent services owned by smaller, independent teams. As a result of these forces, companies are now running hundreds and even thousands of services in production.

The rise of microservices has enabled development teams to scale up and ship code faster, but it has also caused an exponential increase in complexity for operations teams. What was once an in-memory function call is now a cross-continent API request that can fail in unexpected and spectacular ways. What was once a single monitoring dashboard is now a byzantine maze of metrics, logs, and traces. A security model that was once a simple firewall now must protect against a myriad of ever-evolving attack vectors and threats. Finally, what was once a single monolithic service is now hundreds of services built using different technologies and deployed on multiple runtimes: virtual machines (VMs), Kubernetes, serverless platforms, and more.

Operations teams, also known as DevOps and site reliability engineering (SRE), thus face a monumental challenge. In the midst of this complexity, they must harden security, increase ...