Description
A self-contained data lakehouse for analytics experimentation and small-team data platforms: S3-compatible object storage from MinIO, Apache Iceberg-style table layout on raw Parquet, interactive analytics with DuckDB, and visualization with Grafana — plus the Vibe-code agent that can write ETL, SQL and pipelines.
What you get
- MinIO object storage — S3-compatible API for raw Parquet files, with the web console for browsing buckets
- DuckDB analytics sidecar — an in-process columnar OLAP engine that queries Parquet/Iceberg directly over S3, ideal for heavy in-memory analysis without a cluster
- Grafana — dashboards on top of your queries (Postgres/duckdb datasources)
- Room to grow — 640 GB NVMe-class disk and 32 GB RAM for real datasets; architecture points at adding Spark/Trino later
What the Vibe-code agent can do
- Write ETL scripts that land clean Parquet/Iceberg tables in MinIO
- Compose and optimize DuckDB analytical queries (joins, windows, approximations)
- Generate Grafana dashboards from your metrics spec
- Design partitioning, compaction and retention jobs
Server: 8 vCPU · 32 GB RAM · 640 GB SSD (fast NVMe-class), in your chosen region. Start it from My Account → My Servers — ready in about 10–20 minutes.
