# Query every device, site, and reading

Source: https://www.hotdata.dev/industries/iot
Site index: https://www.hotdata.dev/llms.txt

Sensor readings by site, device, and time, next to the records that describe each device.

## How data flows

Data comes from:
- Kafka
- Object storage
- Device registry

Hotdata runs in your VPC (Append history, Latest state, Time buckets), and the data is used by:
- Fleet dashboards
- Alerts
- Product teams

## Common workloads

- **Fleet monitoring.** Latest readings and status for every device, and the ones that went quiet. 5×: Faster than Snowflake, with 10 concurrent clients.
- **Site analytics.** Roll readings up by building, floor, and room over time.
- **Device health.** Catch failing devices early from signal and power trends. [15×: Faster than reading S3 directly.](/blog/object-storage-at-nvme-speed-tested-a-billion-ways#performance-results)
- **Product and field analysis.** Compare hardware versions and installations to see what works best in the field.

## How Hotdata fits

- **Load readings from where they land.** Pull from Kafka topics, or Parquet, CSV, and JSONL in object storage. See [Pull Data](/docs/pull-data).
- **History and current state side by side.** Append readings for full history, and upsert each device's latest state. See [Load modes](/docs/push-data#load-modes).
- **Join readings with device records.** Query readings next to device, site, and installation records, in PostgreSQL, DuckDB, Snowflake, or ClickHouse SQL. See [HotSQL](/docs/sql).
- **Time series in SQL.** Downsample with date_bin and compare periods with window functions. See [Date and time functions](/docs/sql-functions-datetime).
- **Steady latency under load.** With 10 concurrent clients, Hotdata runs 5× faster than Snowflake.

## Get started

The [quick start](/docs/quick-start) gets a table into Hotdata in a few minutes.