Skip to content
audiotale.net
Sign in
NewsNews

Cloudflare Spend Management beta: hard spending limits for Workers and Durable Objects

Cloudflare's new opt-in Spend Management pauses Workers and Durable Objects when a dollar limit is hit; here is how to set it up and what the pause means.

audiotale.net team

· 5 min read

Written by AI from the sources linked below. How we use AI

On this page
  1. Key points
  2. What to do
  3. What happened
  4. Background
  5. How it works
  6. What it means
  7. What we don't know yet
  8. What to watch next
  9. Sources

Cloudflare has opened a beta of Spend Management, an opt-in feature that sets a hard dollar limit on usage-based products and pauses them when the limit is reached. It starts with Workers and Durable Objects and is available to any pay-as-you-go account. Until now, Cloudflare customers could be warned about a growing bill, but not stopped from running it up.

Key points

  • Cloudflare announced Spend Management in beta on 13 November 2026, starting with Workers and Durable Objects, for pay-as-you-go customers.
  • For each enabled product you set a spending limit, a notification email and one or more alert thresholds; at the limit the product is paused automatically.
  • Usage shows in near real time, "within 10 minutes" according to Cloudflare, and Cloudflare says it aims to bring measurement-to-enforcement delay under a minute.
  • It is configured in the dashboard (Manage Account, Billing, Spend limits), with cf billing spend-policies on the CLI, or through a REST API.
  • Earlier in 2026 Cloudflare added email-only Budget alerts (April) and dollar limits for AI Gateway (June); Spend Management is the first account-level pause for Workers.

What to do

  • Workers and Durable Objects customers on pay-as-you-go: open Manage Account, Billing, Spend limits and enable the products you want to protect. Set a limit, a notification email and at least one alert threshold below the limit so you hear before the pause.
  • Think about what pausing means before you set a cap on a production service: once a cap is enforced, Cloudflare says new invocations of that product are rejected. For Workers, Cron Triggers and Queue consumers pause, but static assets still serve.
  • Resuming: raise the limit above the current cycle's spend, disable or delete the blocking policy, or wait for the next billing cycle. Cycles start when the paid subscription starts, not on the calendar month: a subscription that began on 10 October runs roughly 10 October to 10 November.
  • Not on pay-as-you-go? The April Budget alerts page says enterprise contract accounts are not supported there; the Spend Management post says it is available to pay-as-you-go customers. We found no statement about enterprise accounts for Spend Management itself.

What happened

On 13 November 2026 Cloudflare published "Introducing Spend Management: Spending limits and alerts for Cloudflare". It describes an opt-in system launched in beta, starting with Workers and Durable Objects, with more products to follow. It complements, rather than replaces, AI Gateway spend limits, which set per-user and per-model budgets for AI requests routed through a gateway.

For each enabled product the customer sets a spending limit, a notification email and one or more alert thresholds. Emails go out when spend reaches a threshold or the limit. When the limit is reached, a blocking policy applies and the product is paused. The dashboard has an overview of configured products and an audit log of limit changes and cap hits.

Configuration works three ways. In the dashboard, go to Manage Account, Billing, Spend limits. On the command line, cf billing spend-policies supports list, create, get, update and delete, plus audit list and activity list. Through the API, POST /accounts/{id}/billing/spend-policies takes fields including name, product_id (for example "workers"), enabled, threshold_cents (10000 means $100), action ("usage_block") and notification_email.

The post states no separate price for Spend Management.

Background

Cloudflare's usage-based products have been getting more guardrails this year, in steps:

  • April 2026: a Billable Usage dashboard and Budget alerts. The changelog says the dashboard shows daily usage-based costs from the same system that generates the invoice, and that alerts are dollar thresholds that email you when projected monthly spend reaches an amount. The page describes them as notifications only, with no mention of stopping usage. They apply to pay-as-you-go accounts with products such as Workers, R2 and Images.
  • June 2026: spend limits for AI Gateway. Per the changelog and trade coverage, these are cumulative dollar budgets scoped by model, provider or custom metadata such as user or team, with fixed or rolling windows. Requests are blocked when the budget is exceeded, or can be redirected to a fallback model through Dynamic Routes. PPC Land's report says they are available on every plan, and that alerts at the limit were not yet available in that beta.

Spend Management extends the same idea to Workers and Durable Objects, the products most developers build on.

How it works

Cloudflare's post explains the machinery, which matters because it determines how late a cap can be. Usage is metered and aggregated in two stages, about one-minute and about five-minute windows. A Usage Engine partitions processing by account, and a Rating Engine calculates accumulated spend after applying included allowances. Policy evaluation sends alerts through Cloudflare Email Service. Caps propagate asynchronously through Quicksilver, Cloudflare's configuration distribution system, so requests check local state and do not wait on the Rating Engine.

The consequence is a gap between spending and stopping. Cloudflare says usage appears within 10 minutes and that it wants to cut the measurement-to-enforcement delay to under a minute. A cap is therefore a strong safety net, but not an exact ceiling: spend in the minutes before enforcement can overshoot the limit, and the post does not quantify the overshoot, so we do not either.

What it means

For indie developers and small teams, the main benefit is protection from the surprise bill: a runaway loop, a traffic spike or an abused endpoint can now hit a wall at a number you chose. Setting a limit and an earlier alert is a few minutes of work.

For teams running AI agents, the post points toward budgets by application or agent task using metadata tags, so one workload can be paused without affecting others. That is planned, not shipped. For now, the account-level cap is a blunt tool: it pauses the whole product.

For production services, the trade-off is real. A cap protects the budget by turning the product off. A public site that hits its cap will reject new invocations. Teams should decide ahead of time which services are allowed to be paused and set alert thresholds that give a human time to act.

What we don't know yet

  • Whether and when it leaves beta, and its exact pricing, which the post does not mention.
  • Which products come next, beyond "more products are coming".
  • How large the overshoot can be before a cap takes effect.
  • Whether enterprise accounts will get it.

What to watch next

Watch Cloudflare's changelog for R2, D1 and Queues support, for the planned sub-minute enforcement, and for per-application or per-agent budgets using metadata tags.

Sources

SaveStart the discussion

Written by

audiotale.net team

The official account of the people who run audiotale.net: announcements, conversation starters and the newsroom's drafts. Posts written with AI say so. Every other account here is a real person.

Read next