wangfan0840 / kudu-ts

fast time series storage on Apache Kudu

Kudu TS

Kudu TS is a metrics and time series database built on Apache Kudu (incubating). Kudu's focus on real time data ingestion combined with fast scans makes it an ideal platform for storing and querying time series data. Kudu TS includes connectors for popular metrics and alerting tools.

The kernel of Kudu TS lives in the core module, which provides a Java API for efficiently storing and querying metrics in Kudu.

The KTSD module contains a daemon which exposes an HTTP interface to Kudu TS. KTSD's HTTP interface is a drop-in replacement for OpenTSDB's TSD, allowing existing tools with OpenTSDB support to use Kudu TS with minimal effort (such as Grafana).

The argus module contains a Kudu TS backend for the Argus metrics and monitoring platform.

Maturity

Kudu TS is experimental, and does not yet have any stability guarantees.

Contributions

Contributions (through GitHub pull requests) are encouraged and welcomed. If you run into any issues, please do not hesitate to open a GitHub issue.

What is Kudu

Apache Kudu (incubating) is a distributed, columnar storage engine for analytics on fast moving data. Kudu provides real time ingestion with extremely fast scan capability. Kudu's columnar on-disk format makes it extremely efficient at scanning, filtering, and aggregating large time series datasets. The columnar format also provides great compression for time series datasets. The characterstics make Kudu a great platform for a time series database.

License

Kudu TS is distributed under the terms of the Apache License (Version 2.0).

See LICENSE for details.