Open Electricity turns to ClickHouse for faster queries

Open Electricity has deployed ClickHouse to run analytics on its Australian electricity generation data platform, cutting query times across a dataset of about one billion rows.

The company runs an interactive service that aggregates electricity market data from the Australian Energy Market Operator, including generation output, emissions and prices. Records are captured every five minutes and date back to 1999, with the database growing by about 500,000 rows a day.

Universities, government departments, energy businesses and individual developers use the platform to examine the grid's power mix, track output by technology and monitor price movements. As the historical dataset expanded and users ran more complex searches, the analytical performance of Open Electricity's earlier PostgreSQL-based setup declined.

Query times lengthened from near-instant responses to waits of several minutes, limiting the platform's ability to support interactive, near real-time analysis.

Hybrid system

Rather than replace its existing database outright, Open Electricity adopted a hybrid architecture. PostgreSQL continues to handle operational and transactional workloads, while ClickHouse supports large analytical queries and dashboard reporting.

The implementation took about four months and allowed the two systems to operate alongside each other rather than requiring a full migration in a single step.

ClickHouse uses a column-based model that stores data by field rather than by row. This lets queries scan only the parts of the dataset needed for a given analysis, which suits common searches across generation type, emissions and electricity prices.

The new setup also improved storage efficiency. Open Electricity said the historical dataset of around one billion rows can be stored in a relatively small amount of disk space, helping contain infrastructure costs while preserving access to decades of market information.

Faster queries

Performance has improved sharply since the deployment. Queries covering tens of millions of data points now complete in less than a second, while jobs that once took several minutes return almost instantly.

The platform is now used by more than 300 companies and institutions, as well as more than 2,500 individual developers. Broad analytical query loads are also running with p95 latencies below 100ms.

Those gains matter for a user base that includes about 25 universities, multiple state and federal government agencies, energy and grid companies, and financial firms. These groups use the platform to study Australia's energy system and assess the shift toward renewable generation.

The breadth of that audience reflects wider pressure on data systems tracking the energy transition. Australia's electricity market produces continuous streams of operational data, and demand for tools that can interpret long-running datasets quickly has grown as researchers, policymakers and businesses seek clearer views of supply, pricing and emissions trends.

Open Electricity's decision also highlights a broader pattern in enterprise data infrastructure: organisations are keeping established transactional databases in place while adding separate systems tailored to analytical workloads. In this case, the company retained PostgreSQL for day-to-day operational reliability and introduced a specialist analytics database for heavy read queries.

The current architecture also leaves room to broaden the platform's data coverage. One area under consideration is the addition of detailed information on Australia's gas sector, expanding the range of energy market analysis available through the service.

For now, the main effect has been on speed and usability, giving analysts faster access to historical and near real-time electricity market information across a system that continues to grow each day.