Skip to content

ClickHouse Agentic AI & Vector Search in Observability | Observability를 위한 ClickHouse 에이전틱 AI와 벡터 검색

Ken Lee / 이기훈, Solution Architect / 이사 at ClickHouse

이 발표에서는 ClickHouse의 Vector Search와 Agentic AI를 활용해 Observability 데이터를 더 똑똑하게 탐색하는 방법을 소개합니다.

  • 유사 에러 로그 검색: Vector Index(usearch)로 "이 에러와 비슷한 과거 로그" 즉시 탐색
  • 이상 트레이스 탐지: Embedding 기반 정상/비정상 패턴 분류
  • Agentic 워크플로우: MCP를 통해 AI가 직접 ClickHouse를 쿼리하며 근본 원인 분석

본 발표는 ClickHouse 오픈소스로 구현 가능한 기술스택으로 구성하며, ClickHouse Cloud를 통한 라이브 데모로 실제 동작을 확인합니다.


This talk introduces smarter ways to explore Observability data using ClickHouse Vector Search and Agentic AI.

  • Similar Error Log Search: Instantly find "logs similar to this error" with Vector Index (usearch)
  • Anomalous Trace Detection: Classify normal vs abnormal patterns using embeddings
  • Agentic Workflow: AI queries ClickHouse directly via MCP for automated root cause analysis

This presentation is built on a tech stack that can be implemented with ClickHouse open source, with a live demo on ClickHouse Cloud to demonstrate it in action.

Recent videos

YouTube Video: tAmyynjXKMk

Meetups

ClickHouse Delhi/Gurgaon Meetup - Atlys

Presented by, Abhishek Banerjee Senior Data Engineer @ Atlys, at our Delhi/Gurgaon Meetup

YouTube Video: gYe1T0-Yido

Meetups

ClickHouse Delhi/Gurgaon Meetup - Rill Data

Presented by, Nishant Bangarwa Cofounder and Head of Engineering @ Rill Data, at our Delhi/Gurgaon Meetup

YouTube Video: 21LWwIn9y6k

Meetups

ClickHouse Delhi/Gurgaon Meetup - InfoEdge

Presented by, Samyak Jain Senior Data Engineer @ Info Edge, at our Delhi/Gurgaon Meetup

Follow us

XBlueskySlackGithubTelegramMeetupRSS