Back to Projects
Case StudyOffice project

Real-time Event Streaming Pipeline

Built an audit-grade streaming path for service events that favors durability, replayability, and traceable failure handling.

KafkaGoClickHouse

Impact

5M+ events/day · <100ms latency

Source

Private code

Repos

3 linked services

Repository Shape

Event Producers
Private
Audit Consumers
Private
Analytics Sink
Private

Private repositories are represented through architecture notes, impact, and design tradeoffs instead of source links.

Role

Backend implementation across producers, consumers, schema evolution, and ClickHouse ingestion.

Architecture

  • Services publish typed events through Kafka producers with schema validation.
  • Consumer groups separate audit persistence from analytics aggregation so each workload scales independently.
  • ClickHouse stores query-heavy audit history while Kafka retention preserves replay windows.

Highlights

  • Handled multi-million daily event volume with low-latency consumer processing.
  • Separated producer retries from consumer idempotency to avoid duplicate audit records.
  • Added lag and dead-letter monitoring for operational visibility.

Constraints

Office Project

Company-owned repositories cannot be linked publicly, so the page documents architecture and tradeoffs without proprietary code.