Case Study: Cloudflare achieves reliable edge task scheduling with HashiCorp Nomad and Consul

Nomad and Consul Are Being Used at Cloudflare

Cloudflare, a global edge network with hardware in 200+ cities, needed a better way to run internal management workloads without placing them on every machine or forcing teams to manage complex datacenter-specific deployments. The company wanted a dynamic scheduling layer that could improve resource use, isolate workloads, and keep services highly available across its distributed edge.

HashiCorp provided Nomad, integrated with Consul, to run one Nomad server cluster per location and a Nomad client on every machine. Cloudflare used this setup for automated bootstrapping, zero-downtime upgrades, staged job deployments, and service discovery for metrics and logging. The result was a lightweight, low-ops platform that improved reliability after hardware failures, reduced noisy-neighbor risk, and enabled more workloads to move onto the edge; Cloudflare reported that new jobs have been added every quarter.

Cloudflare

Thomas Lefebvre

Tech Lead