Kube32 brings a Kubernetes-style cluster model to ESP32-S3 microcontrollers, letting users manage boards with familiar tools such as kubectl and k9s. One firmware image serves every board: a unit with at least 2 MB of PSRAM can act as control plane, while other ESP32-S3 boards join as workers. Applications run as WebAssembly pods through the wasm3 interpreter, and a macOS/Linux simulator provides a way to build and test multi-node clusters without physical hardware. The project’s central idea is to make small, networked embedded devices manageable as a cluster rather than as isolated gadgets.
The system implements a broad set of Kubernetes-compatible resources and behaviors, including deployments, services, ingress, jobs, configuration, logs, and API watch streams. It adds embedded-focused capabilities: per-pod permissions for GPIO, I2C, UART, displays, and other peripherals; persistent storage on flash or microSD; resource limits and crash-restart policies; and firmware updates with health checks and rollback. Pods can communicate through HTTP, while events can cross nodes using ESP-NOW or UDP. The README demonstrates Tessera, an issue tracker split across four pods: web and search services on the control-plane board, with database storage and a display/keypad interface on a Cardputer worker. Setup involves building and flashing firmware, provisioning Wi-Fi, then deploying workloads through standard Kubernetes manifests.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.