Skip to content

Databases & warehouses

TimescaleDB analytics: ask hypertables and continuous aggregates

TimescaleDB is a time series extension for PostgreSQL, built by the company formerly called Timescale and now Tiger Data, available self-hosted or on its managed Tiger Cloud. Teams store metrics, sensor readings, financial ticks and application events in hypertables, which partition data into time-based chunks, and keep relational tables such as devices or customers in the same database.

TimescaleDB

Available now: Nexatron answers TimescaleDB questions with user-level access that inherits your existing TimescaleDB permissions.

Updated

Can Claude, ChatGPT or Copilot access TimescaleDB data?

Yes. To connect Claude, ChatGPT or Copilot to TimescaleDB, add Nexatron from Claude’s connector directory, or by its MCP server URL in ChatGPT and Copilot. Then connect it in Nexatron. Users sign in with their own TimescaleDB account (or your SSO), and Nexatron inherits their existing permissions.

What TimescaleDB data a question reads

Hypertables and chunks
Tables partitioned automatically by time (and optionally space), such as readings(time, device_id, temperature).
Continuous aggregates
Materialized rollups, such as hourly averages per device, refreshed incrementally by policy.
Columnstore compression
Older chunks converted to compressed columnar form, segmented by a column such as device_id.
Relational tables
Ordinary PostgreSQL tables for devices, sites, customers or instruments, joined to time series.
timescaledb_information views
Hypertables, chunks, continuous_aggregates, jobs and job_stats: sizes, ranges and background job health.
Hyperfunctions
time_bucket, time_weight, percentile approximations and gap filling for time series questions.

TimescaleDB questions people ask

  • What was the hourly average power draw per site over the last 7 days, from the continuous aggregate?
  • Which devices had a temperature excursion above threshold for more than 20 minutes this week?
  • What were daily trading volume and VWAP per instrument last month?
  • Which customers’ API request counts fell more than half week over week?
  • Which continuous aggregate refresh jobs failed in the last 24 hours?
  • Which hypertables are largest, and how much space does compression save on each?

Who asks TimescaleDB questions

Operations and IoT teams
Excursions, uptime and energy use per site or device from sensor hypertables.
Developers and database owners
Chunk sizes, compression ratios and background job failures across hypertables.

How Nexatron connects to TimescaleDB

API and sign-in
The PostgreSQL wire protocol with standard drivers (JDBC, libpq, psycopg), using PostgreSQL roles with SCRAM password authentication or certificates. Tiger Cloud services issue per-role connection credentials.
User-level access
Users sign in with their own TimescaleDB account (or your SSO), and Nexatron inherits their existing permissions. Users only see data they are already authorized to see.
Read-only
Questions only read TimescaleDB; asking a question never changes TimescaleDB data.
Setup
An administrator enables TimescaleDB in Nexatron. Users link their own TimescaleDB account, or sign in through your SSO, the first time they ask. Queries run in TimescaleDB, with no warehouse to load.

TimescaleDB in Claude’s connector directory

Nexatron is listed in Claude’s MCP connector directory, so a Claude user can add it without a custom server URL. Its hosted MCP server answers TimescaleDB questions for Claude, ChatGPT and Microsoft Copilot with user-level access, and the same governed answer as the Nexatron app. It is Nexatron’s own server, not one published by the vendor.

  1. Connect TimescaleDB to your Nexatron workspace.
  2. In Claude, add Nexatron from Claude’s connector directory. In ChatGPT or Copilot, add a custom MCP connector with Nexatron’s server URL.
  3. Sign in with your own Nexatron identity (OAuth 2.1), then ask your TimescaleDB questions.

Server URL for a custom connector: https://api.nexatron.io/api/v1/mcp

TimescaleDB questions and answers

What is a continuous aggregate in TimescaleDB?
A materialized view over a hypertable, grouped with time_bucket, that TimescaleDB refreshes incrementally for new and changed data. Queries on long periods read the rollup instead of raw rows.
Is TimescaleDB just PostgreSQL?
It is a PostgreSQL extension, so standard SQL, joins, indexes, roles and PostgreSQL tools all work. Hypertables, compression, continuous aggregates and hyperfunctions are what the extension adds.
TimescaleDB vs InfluxDB for analytics?
TimescaleDB keeps time series in SQL tables next to relational data, so joins to devices or customers are ordinary SQL. InfluxDB uses its own model of measurements, tags and fields, queried with SQL, InfluxQL or Flux depending on version.
Who makes TimescaleDB now?
Timescale, the company behind TimescaleDB, renamed itself Tiger Data in 2025 and calls its managed service Tiger Cloud. The open-source extension keeps the TimescaleDB name.