THE CRUNCH
Cloudflare has launched eight updates that pull its logs, traces, analytics, alerts, dashboards and data export into a single observability platform, with what it calls simpler and more predictable pricing. The changes let developers investigate issues without first working out which Cloudflare product owns each signal: a spike in 5xx errors, for example, could stem from a Worker, the origin server, or Cloudflare failing to connect to the origin. The company describes this as the first step towards observability becoming a platform-wide capability, with more products and datasets joining over the coming months.
The headline features include a new Logs home combining Workers Observability with Log Explorer, covering datasets such as HTTP events, firewall events, Workers, Containers, R2 and AI Gateway, and Cloudflare Traces in open beta, which shows how individual requests move through security rules, caching, routing, Workers and origin handling. A unified SQL API, also in beta, lets people and AI agents query telemetry across products with one SQL dialect and authentication model, including via the new cf CLI or Cloudflare's Observability MCP server.
On pricing, Cloudflare is consolidating all ingested and stored logs and traces into one Observability subscription. From 1 December 2026 the model applies across all plans, effective on renewal for Enterprise customers. Because logs and traces vary greatly in size, billing is based on ingested and stored volume rather than an event count. The free plan gets 0.5 GB of daily ingestion with 7 days of retention, while paid and Enterprise plans include 50 GB of ingestion and 10 GB-months of storage per billing cycle. Logpush export is also now available on self-serve plans.


