Cloud Backup Platform
Automated, versioned backups for cloud workloads with configurable retention policies and restore workflows.
Problem
Teams needed scheduled, verifiable backups of large object stores without manually managing retention or worrying about silent restore failures.
Architecture
A Celery-based scheduler triggers incremental snapshot jobs against MinIO buckets; metadata and retention rules live in PostgreSQL, with a DRF API exposing backup status and restore points to the frontend.
Challenges
Keeping incremental snapshots consistent under concurrent writes, and making large restores resumable instead of all-or-nothing.
Solution
Introduced checkpointed restore jobs with idempotent chunk processing, so a failed restore resumes from the last verified chunk instead of restarting.