Show HN: Arc – high-throughput time-series warehouse with DuckDB analytics
4 hours ago
- #time-series
- #database
- #performance
- Arc Core is a high-performance time-series data warehouse built on DuckDB, Parquet, and MinIO.
- Currently in alpha release, it is not recommended for production workloads but is stable for development and testing.
- Supports high-performance ingestion via MessagePack binary protocol, InfluxDB Line Protocol, and JSON.
- Features a DuckDB Query Engine for fast analytical queries with SQL.
- Offers distributed storage with MinIO, supporting S3-compatible object storage, local disk, AWS S3, and GCS.
- Capable of importing data from InfluxDB, TimescaleDB, and HTTP endpoints.
- Includes configurable query caching for improved performance.
- Achieves 1.89M records/sec with MessagePack binary protocol.
- Optimal configuration includes 3x CPU cores for workers, native deployment, and MinIO native storage.
- Provides Docker and native deployment options, with native being 2.4x faster.
- Configuration is managed via a centralized arc.conf file with environment variable overrides.
- Requires authentication via Bearer token for API access.
- Supports batch ingestion and is compatible with InfluxDB clients and Telegraf.
- Benchmarked using ClickBench, showing high performance on both AWS and Apple M3 Max hardware.
- Licensed under AGPL-3.0, with commercial support options available.