Skip to main content

5-Minute Edge Quick Start

This guide helps you get MonoTS up and running from source on an edge host. You will build the server, create a table, ingest local sensor data, and query it.

No public Docker image yet

MonoTS is currently in early development. There is no docker pull monots image available today in the public repository. You must use a Rust toolchain to build it from source.

Prerequisites​

  1. Rust toolchain matching rust-toolchain.toml (currently 1.94.1).
  2. Standard build dependencies (CMake, OpenSSL, Protobuf, etc.).

Step 1: Build and start the MonoTS server​

On your edge host or development machine, clone the repository and build the binary.

git clone https://github.com/mono-io/monots.git
cd monots
# Build server and CLI in release mode
make build-host
# Run server with default development config
make run-server

MonoTS Server is now running, listening for gRPC connections on http://127.0.0.1:50051.

SettingValue
gRPC listenhttp://127.0.0.1:50051
Username / passwordadmin / admin

Step 2: Open the local SQL CLI​

In another terminal window, use the MonoTS CLI to interact with your database locally via gRPC.

# In the same monots directory
make run-cli
note

Access is via gRPC, not HTTP curl to port 8080.

Step 3: Model your edge data​

Every table must include a column named exactly time. While TIMESTAMP types are supported, we recommend using BIGINT epoch milliseconds for simplicity with edge sensors.

Run this SQL in the CLI:

CREATE TABLE local_temp_sensor (
  time BIGINT NOT NULL, -- Mandatory for memtable ordering and pruning
  device_id VARCHAR,
  temperature DOUBLE,
  battery_level INT
);

Step 4: Ingest and analyze local data​

Insert data using integer epochs for the time column.

INSERT INTO local_temp_sensor (time, device_id, temperature, battery_level) VALUES
  (1718000000000, 'sensor-a', 23.5, 90),
  (1718000060000, 'sensor-b', 21.8, 88);

Run a local analytical query. Prefer filtering on time so MonoTS can use efficient time-aware pruning.

SELECT device_id, AVG(temperature) AS avg_temp
FROM local_temp_sensor
WHERE time >= 1718000000000
GROUP BY device_id;

Optional: force memtable data to Parquet:

FLUSH TABLE local_temp_sensor;

Next steps​

Your data is now persisted locally on the edge node in optimized Parquet files. Next, configure MonoTS to synchronize this data reliably to the cloud.