Summary
In an October 12, 2026 post, Cloudflare detailed Logpush Transformers, which let teams define a single SQL query to filter records, redact sensitive fields, and reshape log output (adding, renaming, removing, or nesting fields) before logs are exported to each destination.
What changed
Logpush Transformers let a team attach one SQL query (up to 32 KB) to a Logpush job; Cloudflare runs it on every batch to filter records, redact sensitive data, and reshape output so each destination receives data in the form it needs, applied before export. Cloudflare's post states Logpush became available on all self-serve plans and that Transformers reached general availability during Birthday Week.
Why it matters
Pushing filtering and redaction to the edge, before logs leave Cloudflare, reduces downstream storage and SIEM ingest costs and helps teams meet data-handling and privacy requirements by stripping sensitive fields at the source. SQL-based reshaping also removes a common ETL step between Cloudflare logs and diverse destinations, simplifying observability pipelines.
Evidence excerpt
Logpush Transformers let teams use SQL to filter, redact, and reshape logs before they are exported, so each destination receives data in the form it needs.