NATS at the Edge
Resources
Leaf nodes are the backbone of nearly every NATS edge deployment — from vehicles and robots to warehouses, stores, and satellites. But the issues that take down large fleets usually aren’t about how many leaf nodes you run; they’re about what quietly leaks behind them.
In this hands-on RethinkConn workshop, Synadia engineers Borja and Sam Atwood move past “hello world” and show you what really happens on the wire when you connect leaf nodes to NATS — and the patterns that let you scale to millions of devices without overwhelming your hub.
What’s inside
A practical, architecture-first session built around ~20 hands-on labs. Less about connecting a device, more about designing an edge system that stays fast, secure, and stable as it grows.
What you’ll learn
- When to put NATS at the edge — client-at-the-edge vs. running a leaf node on the device itself
- Topologies that fit the edge — clusters, superclusters, and why leaf nodes are uniquely asymmetric
- The three leaf-node boundaries — security, networking, and JetStream — and how to use each one
- Interest propagation — the most overlooked cause of edge scaling problems, and how to control it
- Securing the edge — centralized auth, per-leaf security models, and system-account observability
- JetStream at scale — domains, raft isolation, data locality, and microcell architecture
- Proven scaling patterns — the republish pattern for addressing millions of devices, plus ping/pong and lame-duck tuning
Who it’s for
Architects, platform engineers, and SREs designing edge, IoT, or fleet systems on NATS — anyone who needs real-time command and control, local autonomy during disconnects, and predictable performance at scale.
Includes
Two live demos and a companion labs repo so you can follow along. Bring NATS 2.14, the NATS CLI, and jq.
Keep learning
All workshopsGo deeper with NATS training
Take your skills further with our live, instructor-led courses delivered by our experts to your team of developers or operators.