Raspberry Pi Edge Fleet Tooling & AWS IoT Operations

Overview

This is an edge-device operations project for Raspberry Pi-based field systems deployed at live sports field systems. It focuses on the practical engineering glue needed to run a fleet of physical devices: provisioning new hardware, syncing configuration through the cloud, delivering over-the-air updates, collecting telemetry, and tying device assignments back to the spreadsheets the field teams already use.

Rather than a single polished product, it’s a collection of cooperating pieces — on-device agents, runbooks, CLI tooling, and OTA packaging — that make operating real hardware in the field repeatable and observable.

The Problem

Field devices are messy to operate: each Raspberry Pi needs to be provisioned, configured, kept up to date, and monitored, often from a distance and across many physical locations. Configuration drifts, assignments change from event to event, and pushing code to devices by hand doesn’t scale. The goal was to bridge newer AWS IoT workflows with existing Pi scripts and local config paths into one operational toolchain.

Approach

• Device provisioning — a phased workflow from a fresh Raspberry Pi to a fully managed, OTA-capable device, documented in runbooks.
• Cloud configuration sync — device config is driven through AWS IoT Core using MQTT topics, certificates, and Thing Shadows (desired vs. reported state).
• OTA updates — update artifacts are distributed via S3 with SHA256 verification, orchestrated through AWS IoT Jobs and Job Templates and triggered by MQTT bootstrap commands.
• On-device services — Python services running as systemd units handle GPIO relay control, temperature/humidity sensing, J1939 CAN bus battery data, Victron power data, and local API-style control.
• Telemetry & monitoring — devices emit metrics for monitoring and alerting, so fleet health is observable from a distance.

Assignment Workflow

Field assignments still live in spreadsheets, so the tooling pulls and parses that assignment data and transforms it into AWS IoT shadow update payloads — letting the existing spreadsheet-driven process drive cloud configuration without forcing teams to change how they work day to day.

Fleet CLI

A TypeScript CLI wraps the fleet API for everyday operations: listing devices, patching configuration, setting relay desired-state, and creating rollout jobs. It gives operators a fast, scriptable way to act on the fleet without touching each device individually.

Outcome

A practical operations toolchain — provisioning runbooks, device agents, OTA packaging, telemetry, and CLI tooling — that turns ad-hoc field-device work into a repeatable, monitored workflow bridging Raspberry Pi hardware and AWS IoT.

In Progress

A fuller “Assignments App” intended to replace the spreadsheet workflow is captured as a future direction in the runbooks, and the broader fleet platform (API, device agent, UI, and infrastructure helpers) continues to grow as more of the manual operations are folded into the toolchain.