O11y in One: Exploring ClickHouse as a Unified Telemetry Database
summary
Argues for one unified telemetry store by using ClickHouse to replace separate logs/metrics/traces stacks, claiming it handles petabyte-scale and “trillions of rows” while observability can outstrip app data 50x–100x. Uses ClickHouse MergeTree columnar storage, write-once, read-many workload optimization, materialized views, TTL/tiered storage, and a new JSON column for resource metadata. Shows integrations: OpenTelemetry Collector, Kafka table engine, Grafana, Jaeger; examples of full stacks: ClickStack, Querine, Kroot; notes PromQL and time-series engine are experimental, defaults often use 7–14 day retention, and tiny deployments (≈10 GB) can still work.
AI-written summary · about
chapters (8)
- 0:00Introduction and Overview
- 0:20Understanding Observability and Open Telemetry
- 2:58Challenges with Disparate Telemetry Systems
- 5:27Why ClickHouse for Observability
- 6:38ClickHouse Architecture and Benefits
- 13:34Open Source Solutions and Integrations
- 20:07Scaling and Production Considerations
- 22:42Q&A Session
related talks
36mThe Power of Consolidation: A Unified Stack for Business Intelligence, Securi...
28mColumnstore Revolution
42mClickHouse: what is behind the fastest columnar database
17mBuilding Real-Time Data Lake with ClickHouse
45mClickHouse: what is behind the fastest columnar database
27mFast DIY O11y with OpenSrc Analytics & Visualization
34mClickHouse® - What is Behind the Fastest Columnar Database?
34mThe State of SQL-Based Observability