Website • Documentation • Quick Start

English | 简体中文

CnosDB is a high-performance, high-compression, and easy-to-use open-source distributed time-series database. It is primarily used in fields such as IoT, industrial internet, connected cars, and IT operations. All of the code is open-sourced and available on GitHub.

In its design, we fully utilize the characteristics of time-series data, including structured data, non-transactions, fewer deletions and updates, more writes and less reads, etc. As a result, CnosDB has a number of advantages that set it apart from other time-series databases:

• High performance: CnosDB addresses the issue of time-series data expansion and theoretically supports unlimited time-series data. It supports aggregate queries along the timeline, including queries divided by equal intervals, queries divided by enumeration values of a column, and queries divided by the length of the time interval between adjacent time-series records. It also has caching capabilities for the latest data and the cache space can be configured for fast access to the latest data.
• Easy to use: CnosDB provides clear and simple interfaces, easy configuration options, standard SQL support, seamless integration with third-party tools, and convenient data access functions. It supports schema-less writing mode and supports historical data supplement(including out of order writing).
• Cloud native: CnosDB has a native distributed design, data sharding and partitioning, separation of storage and computing, Quorum mechanism, Kubernetes deployment and complete observability, ensuring final consistency. It can be deployed in public clouds, private clouds, and hybrid clouds. t also supports multi-tenancy and has role-based permission control. The computing and storage nodes support horizontal scaling.

Architecture

Quick Start

Build&Run from source

Support Platform

We support the following platforms, if found to work on a platform not listed, Please report to us.

• Linux x86(x86_64-unknown-linux-gnu)
• Darwin arm(aarch64-apple-darwin)

Requirements

1. Install Rust, You can check official website to download and install
2. Install Cmake

# Debian or Ubuntu
apt-get install cmake
# Arch Linux
pacman -S cmake
# CentOS
yum install cmake
# Fedora
dnf install cmake
# macOS
brew install cmake

3. Install FlatBuffers

# Arch Linux
pacman -S flatbuffers
# Fedora
dnf install flatbuffers
# Ubuntu
snap install flatbuffers
# macOS
brew install flatbuffers

If your system is not listed, you can install FlatBuffers as follows

$ git clone -b v22.9.29 --depth 1 https://github.com/google/flatbuffers.git && cd flatbuffers

# Choose one of the following commands depending on your operating system
$ cmake -G "Unix Makefiles" -DCMAKE_BUILD_TYPE=Release
$ cmake -G "Visual Studio 10" -DCMAKE_BUILD_TYPE=Release
$ cmake -G "Xcode" -DCMAKE_BUILD_TYPE=Release

$ sudo make install

Compile

git clone https://github.com/cnosdb/cnosdb.git && cd cnosdb
make build

Run

Run CnosDB

The following is a single node startup. If you need to start a cluster, see Cluster startup process

./target/debug/cnosdb-meta --id 1 --http-addr 127.0.0.1:21001
curl http://127.0.0.1:21001/init -d '{}'
curl http://127.0.0.1:21001/metrics
./target/debug/cnosdb run --config ./config/config_31001.toml

Run CLI

cargo run --package client --bin cnosdb-cli

Run with Docker

1. Install Docker

2. Start container

docker run --name cnosdb -d  --env cpu=2 --env memory=4 -p 31007:31007 cnosdb/cnosdb:v2.0.1

3. Run a command in the running container

docker exec -it cnosdb sh

4. Run cnosdb-cli

cnosdb-cli

Quit \q Help \? For more details, check basic operation

Write data

• SQL
• influxdb line-protocol
• bulk loading
• telegraf

The following will show an example of using cli to write data by SQL

1. CREATE TABLE

CREATE TABLE air (
    visibility DOUBLE,
    temperature DOUBLE,
    pressure DOUBLE,
    TAGS(station)
);

public ❯ CREATE TABLE air (
    visibility DOUBLE,
    temperature DOUBLE,
    pressure DOUBLE,
    TAGS(station)
);
Query took 0.063 seconds.

2. Insert a row

INSERT INTO air (TIME, station, visibility, temperature, pressure) VALUES
                (1673591597000000000, 'XiaoMaiDao', 56, 69, 77);

public ❯ INSERT INTO air (TIME, station, visibility, temperature, pressure) VALUES
                (1673591597000000000, 'XiaoMaiDao', 56, 69, 77);
+------+
| rows |
+------+
| 1    |
+------+
Query took 0.032 seconds.

3. insert multiple rows

INSERT INTO air (TIME, station, visibility, temperature, pressure) VALUES
                ('2023-01-11 06:40:00', 'XiaoMaiDao', 55, 68, 76),
                ('2023-01-11 07:40:00', 'DaMaiDao', 65, 68, 76);

public ❯ INSERT INTO air (TIME, station, visibility, temperature, pressure) VALUES
                ('2023-01-11 06:40:00', 'XiaoMaiDao', 55, 68, 76),
                ('2023-01-11 07:40:00', 'DaMaiDao', 65, 68, 76);
+------+
| rows |
+------+
| 2    |
+------+
Query took 0.038 seconds.

Query data

• SQL, compatible with SQL standard.
• Prometheus remote read.

The following will show an example of SQL query using cli

-- query table data
SELECT * FROM air;

public ❯ -- query table data
SELECT * FROM air;
+---------------------+------------+------------+-------------+----------+
| time                | station    | visibility | temperature | pressure |
+---------------------+------------+------------+-------------+----------+
| 2023-01-11T06:40:00 | XiaoMaiDao | 55         | 68          | 76       |
| 2023-01-13T06:33:17 | XiaoMaiDao | 56         | 69          | 77       |
| 2023-01-11T07:40:00 | DaMaiDao   | 65         | 68          | 76       |
+---------------------+------------+------------+-------------+----------+
Query took 0.036 seconds.

Connector

• Java: Support JDBC and Java Arrow Flight
• C++: Support ODBC and C++ Arrow Flight
• Go
• Python: Conforms to the Python Data Access Specification (PEP 249)
• Rust

Roadmap

• Click to view Roadmap

Join the community

All developers/users who love time series databases are welcome to participate in the CnosDB User Group. Scan the QR code below and add CC to join the group.

Please check Instructions for joining the group beforehand.

Contributing

Please refer to Contribution Guide to contribute to CnosDB.

Acknowledgement

• CnosDB 2.0 uses Apache Arrow as the memory model.
• CnosDB 2.0's query engine is powered by Apache Arrow DataFusion.
• CnosDB 2.0's bug detection is powered by SQLancer.