GREF Technical Work Report
Overview
This document summarizes the technical work delivered for the Amazon GREF project since its inception in October 2025. The work spans automation, app creation, workflow definition, and API design, with the goal of reducing manual and email-driven processes, improving reliability, and enabling scalable provisioning, configuration, and installation workflows.
Each section below outlines the problem addressed, the solution delivered, and the resulting impact. References to detailed documentation are included where available.
Technical Work Completed
1. Gateway Certificates Retrieval & Distribution Automation System
Overview
Built an automated system to retrieve, package, and distribute gateway certificates and shared configuration files (servertrust and endpoints) as soon as they are made available by the GREF provisioning workflow. The system continuously monitors for new or updated files and ensures warehouse teams automatically receive the latest certificates in their respective regional mailboxes, with no manual intervention.
Why This Was Needed?
Certificate distribution was previously manual and error-prone, leading to delays in gateway configuration and increased risk of missed or inconsistent certificate delivery across regions. Manual intervention resulted in increased operational overhead for warehouse teams.
Key Capabilities
Automated retrieval of gateway certificates and shared configuration files
Warehouse-level segregation to ensure files are delivered only to the correct regional teams
Automated ZIP packaging with size-aware splitting for SES limits
Scheduled processing with retries and failure handling
Centralized logging and delivery status tracking
Scalable and efficient design, reducing manual effort and operational errors
Impact
Eliminated manual certificate distribution steps
Enabled faster and more reliable warehouse configuration
Improved scalability, reliability and operational consistency across all warehouses
References
2. GREF Installs / JSON Creation App (Atria)
Overview
The GREF Installs App in Atria streamlines the creation and management of device configuration JSONs for sensors and gateways. It replaces manual, Excel-based processes with a collaborative, standardized, and time-efficient workflow. These JSONs are a critical input to the GREF workflow and are handed over to the GREF team post-installation for downstream processing.
Why It Was Needed?
Manual JSON creation was error-prone and slow
Excel-based collaboration caused delays and inconsistencies
Multiple stakeholders needed a way to incrementally update JSONs across install stages
Key Capabilities
Create and validate device configuration JSONs
Auto-save (after 30 seconds of inactivity), manual save, and data persistence across refreshes and offline scenarios
Standardized metadata handling (DevEUIs, pulse rates, locations, floors, etc.)
Support multi-stage collaboration:
Add metering numbers after site surveys/scoping
Add device EUIs after procurement
Add installation-specific data post-install (e.g., pulse rates, locations, floors, etc.)
Ability to add new GREF sites directly via the app
Option to create complete JSONs in one step immediately after installation
Impact
Automated and scalable JSON creation workflow
Significant time savings and improved efficiency
Reduced downstream configuration errors
Consistent, reliable JSONs across sites
Eliminates manual Excel-based workflows
References
https://wiki.inside.milvian.group/wiki/spaces/SE/pages/612794382
3. Device & Site Health Check APIs (Email Bot Replacement)
Overview
Proposal to replace GREF’s existing SES-based email bot with REST APIs that directly query GREF backend services (e.g., AWS Timestream) to provide device and site health verification during installations. The APIs will be designed to mirror existing email bot reports, enabling a smooth transition for installers and future UI consumers. These endpoints will be consumed by an Atria applet, enabling faster and more reliable on-site verification and paving the way for eventual email bot deprecation.
Why It Was Needed?
Email-based workflow was slow, prone to retries, and hard to scale
Manual email handling caused delays and occasional non-responses
Installers needed a reliable, real-time verification mechanism while on site
Key Capabilities
Device Information API: Returns metadata, health status, and recent telemetry for individual devices
Site Report API: Provides site-level summaries and per-device health details
Unified Schema: Supports both pulse counter and Subeca devices
Time-based Health Status: Healthy, Warning, Critical based on last telemetry
API-First Architecture: Enables future UI and automation integration. APIs provide direct access to backend data.
Impact
Reduced operational dependence on email workflow
Improved installation confidence, verification speed and reliability
Laid foundation for scalable, API-driven GREF operations
References:
4. GREF Provisioning, Configuration & Installations Workflow
Overview
Defined the end-to-end GREF provisioning, configuration, and installation workflow as a clear visual flowchart. The workflow captures ownership and handoffs across all stages, including site surveys, procurement, hardware assignment, provisioning, gateway configuration, shipping/logistics, and on-site installation.
Why It Was Needed?
There was no single, clear visual representation of the full GREF workflow
Ownership across teams and stages was unclear
Lack of clarity caused ambiguity during different phases of the GREF workflow
Impact
Established a shared source of truth across the different teams
Reduced ambiguity during execution of GREF workflows
Streamlined onboarding for new stakeholders
Enabled more consistent, reliable, and scalable execution
References / Flowchart
5. GREF Box Auto-Notification Setup
Overview
Enabled automatic notifications to GREF stakeholders whenever Milvian hardware or installation teams upload JSON files to GREF Box. Notifications are triggered for uploads to both the Device Metadata and Device Config JSONs folders. This automation was implemented by the GREF team following discussions, with testing and validation performed in collaboration with Milvian.
Why It Was Needed?
GREF stakeholders previously had to be manually looped in on every JSON upload
Manual checks were time-consuming, prone to errors, and easy to forget
Delays slowed provisioning and installation handoffs
Key Capabilities
Automatic notifications triggered on file uploads
Covers Device Metadata and Device Config JSONs (i.e., install JSONs) folders
Coverage for both Device Metadata and Device Config JSONs (i.e., install JSONs) folders
Eliminates reliance on manual confirmations
Impact
Eliminates manual follow-ups
Improved visibility into JSON availability and install readiness
Speeds up handoff between teams
Upcoming Work
6. Atria Migration & GREF Applet Development (Updated 2026-03-10)
Note: Original plan, no longer being pursued due to GREF security and governance concerns.
Original Plan: Develop an Atria applet that consumes the Device and Site Health Check APIs (section 3) to provide installers with real-time, on-site verification of device connectivity, health, and site readiness.
Overview
We are migrating the Atria platform to the GREF AWS account, porting it to operate within GREF’s infrastructure, Git repositories, CI/CD pipelines, tooling, and engineering standards. Post-migration, a GREF-specific Atria applet, tentatively titled “GREF Device & Site Health Report”, will be deployed to this instance. It will deliver site- and device-level telemetry from GREF APIs to installers on-site, enabling real-time verification of connectivity and site readiness. This SoW-based engagement reduces reliance on email workflows, improves installation efficiency, and provides a direct revenue opportunity with potential for future work within Amazon/GREF environments.
Why It’s Needed?
Address GREF security and governance requirements
Replace slow, error-prone email workflows with a reliable solution
Enable installers to perform real-time, on-site verification of device and site health
Early detection of issues is critical for efficient site installations
Key Capabilities
Display device-level and site-level telemetry directly in Atria
Highlight configuration or connectivity issues in real time
UI aligned with familiar workflows to minimize retraining
Expected Impact
Reduces reliance on email-based workflows
Improves installation confidence, verification speed, and reliability
Provides a robust, scalable solution that directly consumes GREF API data, eliminating SES-related roadblocks
References: