ClickHouse
ClickHouse is a column-oriented database management system designed for online analytical processing (OLAP) at scale. In Ilum, ClickHouse is available as an optional module for teams that need a dedicated low-latency analytics store alongside the lakehouse.
ClickHouse is disabled by default in Ilum.
When to use ClickHouse
ClickHouse complements Ilum's primary execution engines (Spark, Trino, DuckDB, Flink) in scenarios where:
- The workload demands sub-second response times on pre-aggregated, structured data.
- The data fits a columnar OLAP model and benefits from ClickHouse's specialized storage and compression.
- Downstream consumers (dashboards, real-time APIs) need a high-concurrency query engine separate from the lakehouse.
For ad-hoc analytics and federation across the lakehouse, त्रिगुण remains the right primary engine. For heavy ETL and ML workloads, अपाचे स्पार्क is the right choice. ClickHouse is an additional option, not a replacement.
Integration with Ilum
When enabled, ClickHouse runs as a sub-chart inside the Ilum Helm umbrella. It uses ZooKeeper for coordination and shares Ilum's storage and identity stack.
Data can flow into ClickHouse through:
- Spark batch writes from the lakehouse for nightly aggregation tables.
- Flink streaming sinks for real-time materialization of event streams.
- Direct ingestion from Kafka topics for low-latency event capture.
संरूपण
Enable ClickHouse through the umbrella Helm chart:
clickhouse:
सक्षम : सच्चा
Refer to the official ClickHouse Helm documentation for the full set of values, including replica counts, persistence configuration, and resource limits.