Observability pricing (from December 1, 2026)
Cloudflare’s unified pricing for logs and traces: Workers Logs, Workers Traces, Containers logs, R2 data access logs and AI Gateway logs move to per-GB ingestion and GB-month storage on December 1, 2026.
- Products
- Workers LogsWorkersContainersAI Gateway
- Source
- Official docsCloudflare Docs
Key points
- Free: 0.5 GB ingestion per day with seven-day retention; at the daily limit Cloudflare stops ingesting until 00:00 UTC. Paid: 50 GB ingestion and 10 GB-month storage per billing cycle, then $0.25 per GB and $0.10 per GB-month. Paid usage continues automatically past the allowance, and allowances do not roll over.
- Ingestion counts the event body plus default and custom attributes before compression; sampling before storage reduces both meters. The docs’ example of 75 GB ingested and 20 GB-month stored adds $7.05. Unsampled security datasets are $1 per GB with 30-day retention.
- Until December 1 the Workers Logs table still applies: Free 200,000 events a day kept 3 days; Paid 20 million a month kept 7 days, then $0.60 per million. Usage before December 1 is not charged retroactively; Enterprise moves at contract renewal.
Original sources
Last updated
Observability pricing (from December 1, 2026)
Open original sourceCloudflare’s unified pricing for logs and traces: Workers Logs, Workers Traces, Containers logs, R2 data access logs and AI Gateway logs move to per-GB ingestion and GB-month storage on December 1, 2026.
Cloudflare Blog: announces the unified observability pricing and opens Logpush to self-serve plans, with 25 GB included per month for exports to Cloudflare or external destinations ($0.03 or $0.10 per GB beyond) and 1 GB for Transformers ($0.04 per GB beyond).
Open original source
More to read
Why hard spending caps take time: Ashley Peacock on billing latency
Ashley Peacock (@ashleypeacock), who works on Durable Objects at Cloudflare, explains from experience why near-real-time spending caps are hard to build, says hard billing caps are coming soon, and suggests budget alerts plus issue alerts in the meantime.
babynarc: a Durable Objects alarm loop and a surprise bill that grew from $300 to $500.
@babynarc says an alarm loop in an old Durable Object he thought was inactive left him badly burned: he watched a surprise $300 bill grow to $500 even after killing it. Replying to Cloudflare's Dane Knecht and Kenton Varda, he suggests prepending budget-alert details to every `cf` result by default.
A static site that sent every request through the Worker: check run_worker_first
gm365 got repeated Workers usage-alert emails (up to 92%) on a static site and traced them to an AI-written "run_worker_first": true in the config.