Using ClickHouse for logs & traces

Keep every log and trace, and still query them fast

If you're sampling traces, shortening retention or rolling up data to keep your observability bill in check, you're losing the detail you need when something breaks.

ClickHouse is a columnar database built for fast analytics on high-cardinality data. Shopify, Anthropic and Character.ai use it to run observability at petabyte scale. ClickStack, the open source observability stack built on ClickHouse, packages it with OpenTelemetry and a UI for search, dashboards and alerts.

What the session covers:

Join Mike Shi, Principal Product Manager of Observability for a technical walkthrough of storing, searching and analyzing logs and traces in ClickHouse.

• How columnar storage and parallel queries handle high-cardinality OTel data, with 10–30x compression
• Sending logs and traces with OpenTelemetry, using the built-in schemas or your own event format
• Live demo: searching logs, switching to SQL, and following a trace from an error to its root cause
• How ClickStack links logs, traces and metrics at the database layer, with no extra pipelines
• Long-term retention on object storage for under a cent per GB, with no sampling and no rollups
• When to run ClickStack yourself and when to use Managed ClickStack on ClickHouse Cloud
• Live Q&A with the ClickHouse team

Who should attend: SREs, platform and DevOps engineers, and engineering leaders who are hitting cost, scale or retention limits with their current logging and tracing setup, including teams comparing alternatives to Elastic, Datadog, Splunk or OpenSearch.