BTrDB
What it is
BTrDB is an open-source, self-hosted time-series database for dense, streaming telemetry, including high-precision sensor data. It partitions data by time, groups deployments in clusters, and records changes to a stream as new versions. Queries can be pinned to a particular version. Streams have unique UUIDs and can be organized into collections with tags and annotations. The documented Python client connects to a BTrDB server and supports stream lookup and retrieval with nanosecond timestamps; data can also be inserted through Python, PyArrow tables, or pandas DataFrames. Statistical queries can return minimum, maximum, mean, count, and standard deviation without reading every underlying point. StreamSets can be converted to pandas DataFrames or CSV, while Arrow-based tabular methods require commercial customers with Arrow-enabled servers. The documentation includes Dash and Plotly examples and support for passing connections and data objects to Ray workers. Access uses TLS on port 4411 and API-key authentication. The server repository identifies the MIT license. The documentation reports performance and compression measurements for a four-node cluster and high-precision streams.
For and against
For
- Open-source server with self-hosted deployment
- Versioned streams and nanosecond timestamp queries
- Statistical summaries without reading every underlying point
- Python, PyArrow, and pandas insertion methods
Against
- Arrow tabular methods require commercial customers and enabled servers
- Data requests are fully materialized
- TLS is required for API-key authentication
The verdict
BTrDB is aimed at dense time-series workloads, with versioned streams and a Python-centered workflow. Check the commercial restriction on Arrow-based tabular methods and the fully materialized request behavior against your data-processing needs.
Who it suits
BTrDB suits developers and operators working with dense telemetry who can deploy and manage a self-hosted database. Its Python client, tabular conversions, and plotting examples may fit teams that work with pandas, Dash, or Plotly.
Plans and prices
BTrDB
No plan or usage limits stated on the opened pages
Read 11 Oct 2026 · github.com
On the record
- Aggregates
- StatPoint queries can return min, max, mean, count, and standard deviation, with query time proportional to the number of results.btrdb.readthedocs.io · read 4 Oct 2026
- Client library
- The documentation provides Python access to BTrDB through the btrdb package.btrdb.readthedocs.io · read 4 Oct 2026
- Commercial feature limit
- The documentation says native tabular data formats are enabled for commercial customers.btrdb.readthedocs.io · read 11 Oct 2026
- Compression
- The documentation reports 2.93x compression for high-precision nanosecond streams.btrdb.readthedocs.io · read 11 Oct 2026
- Data formats
- StreamSets can be filtered and converted to formats including pandas DataFrames and CSV files.btrdb.readthedocs.io · read 4 Oct 2026
- Data integrations
- The documentation describes StreamSet conversions to pandas DataFrames and CSV, and Arrow-enabled methods for tabular data.btrdb.readthedocs.io · read 11 Oct 2026
- Data request behavior
- The docs state that data requests are fully materialized and that generator-based processing is planned for a future release.btrdb.readthedocs.io · read 11 Oct 2026
- Install
- The maker recommends installing btrdb-python with pip and also documents installation with Anaconda.btrdb.readthedocs.io · read 4 Oct 2026
- Installation
- The maker recommends installing btrdb-python with pip and offers data-science and all-optional-dependency extras.btrdb.readthedocs.io · read 11 Oct 2026
- Integrations
- The Python documentation includes examples using Dash and Plotly to plot data queried from BTrDB.btrdb.readthedocs.io · read 4 Oct 2026
- License
- The BTrDB server repository identifies its license as MIT.github.com · read 11 Oct 2026
- Nanosecond timestamps
- BTrDB’s example client uses nanosecond timestamps to query time-series values.btrdb.readthedocs.io · read 4 Oct 2026
- Parallel processing
- The client documents a serializer for passing BTrDB connections, streams, and StreamSets to Ray workers.btrdb.readthedocs.io · read 4 Oct 2026
- Performance
- The documentation reports 53 million inserted values per second and 119 million queried values per second on a measured four-node cluster of large EC2 nodes.btrdb.readthedocs.io · read 11 Oct 2026
- Product
- BTrDB is a time-partitioned, version-annotated clustered time-series database for high-density univariate data.btrdb.readthedocs.io · read 11 Oct 2026
- Python client
- The documented btrdb-python library connects to a BTrDB server and supports stream lookup and time-series retrieval.btrdb.readthedocs.io · read 11 Oct 2026
- Query summaries
- Statistical points can provide min, max, mean, count, and standard deviation without reading every underlying point.btrdb.readthedocs.io · read 11 Oct 2026
- Security
- The documentation describes TLS access on port 4411 and API-key authentication, which requires the TLS port.btrdb.readthedocs.io · read 4 Oct 2026
