Abstract:

In prototype telemetry systems, direct storage of data as CSV files, JSON files, or raw sample dumps is useful during early testing, but it complicates session lifecycle management, file integrity control, event history tracking, and separation between clients and the database layer. The aim of this work was to design and implement a local telemetry backend on Raspberry Pi 5 as an intermediary layer between API clients, a PostgreSQL database, and a local file repository. The prototype was implemented using Python, FastAPI, Uvicorn, and PostgreSQL. A hybrid persistence model was adopted, in which the relational database stores metadata, statuses, relationships, and events, while measurement files remain in the file system. The central element of the solution is a telemetry session model that organizes the process from session initialization, through file upload and SHA-256 integrity checking, to session finalization and preparation of data for further processing. Functional verification covered 30 CSV files divided into three size groups and five negative or boundary cases. All 30 valid uploads satisfied the adopted functional criteria, while the five negative and boundary cases resulted in the expected request rejection or in data not being marked as ready for further analysis.