Demos
GlassFlow comes with ready-to-run demos that showcase different pipeline capabilities. Each demo uses the GlassFlow CLI to spin up a local environment with all required infrastructure.
Available Demos
-
Kafka Deduplication — Create a pipeline via the UI that deduplicates Kafka events before writing to ClickHouse. A good starting point to understand the core GlassFlow workflow.
-
Fraud Detection — Detect brute-force login attacks by combining Filter and Deduplication in a single pipeline. Uses synthetic login events and SQL-based fraud queries on clean data in ClickHouse.
-
OpenTelemetry Traces — Ingest OpenTelemetry traces into ClickHouse with deduplication and PII masking. Uses the OTLP source, Deduplication on a composite
trace_id+span_idkey, and a stateless transformation for PII redaction, then visualizes traces in HyperDX. -
5G Telemetry: Replay real 5G RAN signal traces as OpenTelemetry metrics from a redundant collector pair. Uses the OTLP source, Filter, Deduplication, and a Stateless Transformation to remove duplicates and normalize cell IDs before being ingested to ClickHouse, and visualized in Grafana.
Prerequisites
All demos require:
- Docker (or compatible runtime like Docker Desktop, OrbStack, Colima, or Podman)
- Helm — install via Helm docs or run:
brew install helm - GlassFlow CLI — see Install with CLI