Event Analytics to ClickHouse

Event Analytics to ClickHouse

This template shows the ClickHouse node on two paths: an ingest path that writes a shaped event row into a ClickHouse table, and a scheduled report path that runs a raw aggregate query. It mirrors the DataTable operation set (insert / find / count / update / remove / upsert) against an external OLAP database, plus a raw query escape hatch.

Ingest path

1. EventInput provides a sample event payload
2. ShapeRow maps the input into the table's columns with the Set node
3. InsertEvent writes the row into the events table via the ClickHouse insert operation

Report path

1. DailySchedule fires once a day with Cron
2. DailyReport runs a raw SQL aggregate (SELECT event, count() ...) with the ClickHouse query operation
3. ReportOutput exposes the aggregated rows

Use cases

• Streaming product or app events into ClickHouse for analytics
• Scheduled rollups and daily/weekly reporting over high-volume event data
• Replacing hand-written HTTP calls to the ClickHouse interface with a visual node

Setup

Create a ClickHouse credential (host, port, username/password, database, secure) and select it in both ClickHouse nodes. Point clickhouseTable at a table that exists in your database — back it with a ReplacingMergeTree engine if you plan to use upsert.

Notes

ClickHouse update and remove run as asynchronous mutations, so prefer batched inserts for high-volume ingest. Raw query returns rows for SELECT statements and a summary for everything else.