# Cloudflare turns its edge, Workers and origin path into one trace

Cloudflare has launched end-to-end request tracing, a unified SQL query layer and common log/trace pricing as part of a broader observability rebuild that spans its network, Workers and customer origins.

The practical change is that debugging a Cloudflare-backed application no longer has to stop at the Worker boundary: one trace can follow security, cache, routing, Worker and origin handling, while logs and traces move toward one query and pricing model.

- Status: Active
- Published: 2026-10-04T09:22:27+13:00
- Updated: 2026-10-04T09:22:27+13:00
- Categories: Web Development, Cloud & Infrastructure, Cloud Platforms, Developer Tools, Observability
- Tags: Cloudflare, distributed tracing, logs, Observability, OpenTelemetry, SQL
- Canonical HTML: https://beyondthe.news/dossiers/cloudflare-unified-observability-traces-sql-pricing

## What changed

On October 2, 2026 Cloudflare announced eight observability updates spanning logs, traces, analytics, alerts, dashboards and telemetry export. Cloudflare Traces is in open beta and records supported request-processing steps as spans, including security rules, transformations, cache decisions, routing, Workers and origin handling. A unified SQL API in beta provides one query surface across Cloudflare data, while Workers logs and OpenTelemetry traces can now appear in Custom Dashboards. Cloudflare is also replacing separate observability charging with a common ingestion-and-storage model beginning December 1, 2026.

## Why it matters

Cloudflare sits in front of applications as well as running code and state for many of them, so failures often cross product boundaries. Request-level traces reduce the need to reconstruct a path from unrelated logs, and the SQL layer makes the same telemetry easier to query from software and agents. The pricing change matters operationally too: teams can model logs and traces under one ingestion/storage scheme rather than treating each signal as a separate product.

## One trace can cross Cloudflare and the origin

Cloudflare Traces records supported edge processing as spans and can continue through Workers and downstream services using trace context. The aim is to show where a request spent time and which Cloudflare decisions affected it.

## SQL becomes the common query surface

The new unified SQL API is intended to query logs, analytics, traces and security data through one dialect and authentication model. Cloudflare is also exposing observability data through its CLI and agent-oriented tooling.

## Logs and traces get a common cost model

Beginning December 1, Cloudflare says logs and traces will share account-level ingestion and storage allowances. Paid and Enterprise plans include 50 GB of ingestion and 10 GB-month of storage per billing cycle, with additional usage priced by volume; unsampled security datasets have separate pricing.

## Key details

- The observability bundle launched October 2, 2026.
- Cloudflare Traces is in open beta and follows supported request processing through edge controls, Workers and origin handling.
- A unified SQL API is in beta.
- Workers logs and OpenTelemetry traces can be used in Custom Dashboards.
- Cloudflare is adding custom alerts and broader self-serve Logpush access.
- Unified logs/traces pricing begins December 1, 2026.
- The current pricing documentation says queries, dashboards, alerts and analytics do not incur additional charge.

## Builder takeaways

- Use traces to distinguish failures in Worker code from cache, routing, security-rule or origin behaviour before adding more application logging.
- Preserve W3C trace context into downstream services if you want the Cloudflare trace to connect to the rest of your stack.
- Recalculate telemetry retention and sampling before the December pricing transition rather than assuming today's Workers event model remains the cost boundary.
- The unified SQL surface may be useful for automated incident investigation, but keep external telemetry exports where vendor independence or longer retention matters.

## What to watch

- General availability and limits for Cloudflare Traces and the unified SQL API.
- How reliably traces cross origins and non-Cloudflare downstream services in real deployments.
- Final December pricing behaviour for existing Enterprise contracts and high-volume security datasets.
- Whether agent/MCP observability tooling becomes a practical incident-response interface rather than a demo layer.

## Uncertainties

- Several components are still beta or open beta.
- End-to-end visibility depends on supported spans and downstream trace-context propagation.
- The operational cost impact varies substantially with log volume, sampling and retention.

## Sources

- [8 major updates to Cloudflare Observability](https://blog.cloudflare.com/one-observability-platform/) — Cloudflare · primary · 2026-10-02T00:00:00+13:00. Primary bundle announcement.
- [Introducing Cloudflare Traces: follow requests through our entire platform](https://blog.cloudflare.com/cloudflare-tracing/) — Cloudflare · primary · 2026-10-02T00:00:00+13:00. Tracing architecture and request-path details.
- [Observability pricing](https://developers.cloudflare.com/observability/pricing/) — Cloudflare Developers · primary documentation · 2026-10-02T00:00:00+13:00. Current December 1 pricing model and allowances.

