Base
Hardened build
minimal ports, security patches applied at build time
Arc Time-Series Database
Azure
Databases
Arc, a high throughput time series and analytical database that ingests millions of points a second and answers analytical SQL over open Apache Parquet files.
Overview
Arc is an open source time series and analytical database from Basekick Labs. It pairs a high throughput ingestion pipeline with the DuckDB analytical query engine and stores every record as ordinary Apache Parquet files, so the data you write stays in a portable, open format that DuckDB, pandas, polars or any other Parquet reader can open directly. There is no proprietary on disk format to migrate out of later.
Arc runs as one statically linked binary. Ingestion, storage, background compaction, SQL query and backup all live in a single process: no JVM, no Python runtime, no separate metadata service and no cluster to operate. It speaks the InfluxDB line protocol, so Telegraf and existing InfluxDB 1.x clients write to it without code changes, and it also accepts MessagePack columnar and MQTT ingestion. Queries are standard analytical SQL with window functions, common table expressions, joins and aggregations, returned as JSON or Apache Arrow, and Prometheus format metrics are exposed for monitoring. Background compaction merges small Parquet files automatically so queries stay fast and storage stays cheap.
Arc is a headless database with no web interface: you operate it over its HTTPS API with curl, the Python SDK, a Grafana data source, or any line protocol client.
Why the cloudimg image
Arc ships as a bare binary with token authentication that has to be turned on, TLS that has to be configured, telemetry that is on by default, and a benchmark tuned sample configuration that buffers five million records in memory and would exhaust a small virtual machine. cloudimg closes every one of those gaps before the machine is ever reachable. Token authentication is enforced on every query, write and administrative route, and the only endpoints that answer without a token are the two liveness probes and the two aggregate monitoring endpoints, which carry no queries, no measurement names, no file paths and no credentials. A unique admin API token and a unique TLS certificate are generated on the first boot of every machine and written to a file only root can read, so no credential is baked into the image and no two deployments share a secret; the database is held closed until that first boot completes. Arc terminates TLS itself, so tokens never cross the network in clear text. Usage telemetry is switched off, no licence key is configured, and the appliance never contacts the vendor, so it runs fully air gapped. Time series data is written to a dedicated data volume that is captured into the image and re provisioned on every machine, so your store is sized, snapshotted and resized independently of the operating system disk. The memory limits are sized for the recommended machine rather than for a benchmark. The base is fully patched with unattended security updates left enabled, and every deployment is paired with a step by step deploy guide verified against this exact build and backed by 24/7 cloudimg support.
Common uses
- Store observability and infrastructure metrics and query them with analytical SQL
- Capture industrial and IoT sensor telemetry over the InfluxDB line protocol
- Keep product analytics and event data in open Parquet files you can read with any tool