Cloudflare Pipelines as a Logpush destination
Key Points
- Logpush destination: Pipelines
- Store logs as Parquet or Iceberg in R2
- Transform logs in-flight with Pipelines SQL
Summary
Cloudflare Logpush can now send logs directly to Pipelines, enabling ingestion, transformation, and storage in R2 as Parquet files or Apache Iceberg tables managed by R2 Data Catalog. This reduces storage footprint and improves query performance with R2 SQL or any engine that supports Iceberg/Parquet.
Key Points
- Logpush → Pipelines: send Cloudflare logs directly to Pipelines as a Logpush destination.
- Storage formats: write compact Parquet files or managed Apache Iceberg tables in R2 (R2 Data Catalog).
- In-flight transforms: Pipelines SQL runs per-record to drop fields, redact sensitive values, derive columns, convert timestamps, hash values, extract JSON, apply regex, and use conditional logic.
- Querying: query logs instantly with R2 SQL or any Iceberg/Parquet-compatible query engine.
- Operational benefit: lower storage footprint and faster analytics at scale.
Getting started
- Configure Pipelines as a Logpush destination using the Cloudflare documentation (Enable Cloudflare Pipelines).