Cloudflare Docs
Workers
Edit this page on GitHub
Set theme to dark (⇧+D)

Limits

​​ Account plan limits

FeatureWorkers FreeWorkers Paid (Bundled, Unbound) and Standard
Subrequests50/request50/request (Bundled),
1000/request (Unbound, Standard)
Simultaneous outgoing
connections/request
66
Environment variables64/Worker128/Worker
Environment variable
size
5 KB5 KB
Worker size1 MB10 MB
Worker startup time400 ms400 ms
Number of Workers100500
Number of Cron Triggers
per account
5250

​​ Request limits

URLs have a limit of 16 KB.

Request headers observe a total limit of 32 KB, but each header is limited to 16 KB.

Cloudflare has network-wide limits on the request body size. This limit is tied to your Cloudflare account’s plan, which is separate from your Workers plan. When the request body size of your POST/PUT/PATCH requests exceed your plan’s limit, the request is rejected with a (413) Request entity too large error.

Cloudflare Enterprise customers may contact their account team or Cloudflare Support to have a request body limit beyond 500 MB.

Cloudflare PlanMaximum body size
Free100 MB
Pro100 MB
Business200 MB
Enterprise500 MB (by default)

​​ Response limits

Cloudflare does not enforce response limits, but cache limits for Cloudflare’s CDN are observed. Maximum file size is 512 MB for Free, Pro, and Business customers and 5 GB for Enterprise customers.


​​ Worker limits

FeatureFreeBundled usage modelUnbound and Standard usage model
Request100,000 requests/day
1000 requests/min
nonenone
Worker memory128 MB128 MB128 MB
CPU time10 ms50 ms HTTP request
50 ms Cron Trigger
30 s HTTP request
15 min Cron Trigger
15 min Queue Consumer
DurationNonenonenone

​​ Duration

Duration is a measurement of wall-clock time — the total amount of time from the start to end of an invocation of a Worker. There is no hard limit on the duration of a Worker. As long as the client that sent the request remains connected, the Worker can continue processing, making subrequests, and setting timeouts on behalf of that request. When the client disconnects, all tasks associated with that client request are canceled. Use event.waitUntil() to delay cancellation for another 30 seconds or until the promise passed to waitUntil() completes.

​​ CPU time

CPU time is the amount of time the CPU actually spends doing work, during a given request. Most Workers requests consume less than a millisecond of CPU time. It is rare to find normally operating Workers that exceed the CPU time limit.


​​ Cache API limits

FeatureWorkers FreeBundledUnboundStandard
Max object size512 MB512 MB512 MB512 MB
Calls/request50501,0001,000
Storage/request5 GB5 GB5 GB5 GB
  • 50 total put(), match(), or delete() calls per-request, using the same quota as fetch().

  • 5 GB total put() per request.


​​ Request

Workers automatically scale onto thousands of Cloudflare global network servers around the world. There is no general limit to the number of requests per second Workers can handle.

Cloudflare’s abuse protection methods do not affect well-intentioned traffic. However, if you send many thousands of requests per second from a small number of client IP addresses, you can inadvertently trigger Cloudflare’s abuse protection. If you expect to receive 1015 errors in response to traffic or expect your application to incur these errors, contact your Cloudflare account team to increase your limit.

The burst rate and daily request limits apply at the account level, meaning that requests on your *.workers.dev subdomain count toward the same limit as your zones. Upgrade to a Workers Paid plan to automatically lift these limits.

​​ Burst rate

Accounts using the Workers Free plan are subject to a burst rate limit of 1,000 requests per minute. Users visiting a rate limited site will receive a Cloudflare 1015 error page. However if you are calling your Worker programmatically, you can detect the rate limit page and handle it yourself by looking for HTTP status code 429.

Workers being rate-limited by Anti-Abuse Protection are also visible from the Cloudflare dashboard:

  1. Log in to the Cloudflare dashboard and select your account and your website.
  2. Select Security > Events > scroll to Activity log.
  3. Review the log for a Web Application Firewall block event with a ruleID of worker.

​​ Daily request

Accounts using the Workers Free plan are subject to a daily request limit of 100,000 requests. Free plan daily requests counts reset at midnight UTC. A Worker that fails as a result of daily request limit errors can be configured by toggling its corresponding route in two modes: 1) Fail open and 2) Fail closed.

​​ Fail open

Routes in fail open mode will bypass the failing Worker and prevent it from operating on incoming traffic. Incoming requests will behave as if there was no Worker.

​​ Fail closed

Routes in fail closed mode will display a Cloudflare 1027 error page to visitors, signifying the Worker has been temporarily disabled. Cloudflare recommends this option if your Worker is performing security related tasks.


​​ Memory

Only one Workers instance runs on each of the many global Cloudflare global network servers. Each Workers instance can consume up to 128 MB of memory. Use global variables to persist data between requests on individual nodes. Note however, that nodes are occasionally evicted from memory.

If a Worker processes a request that pushes the Worker over the 128 MB limit, the Cloudflare Workers runtime may cancel one or more requests. To view these errors, as well as CPU limit overages:

  1. Log in to the Cloudflare dashboard and select your account.
  2. Select Workers & Pages and in Overview, select the Worker you would like to investigate.
  3. Under Metrics, select Errors > Invocation Statuses and examine Exceeded Memory.

Use the TransformStream API to stream responses if you are concerned about memory usage. This avoids loading an entire response into memory.


​​ Subrequests

A subrequest is any request that a Worker makes to either Internet resources using the Fetch API or requests to other Cloudflare services like R2, KV, or D1.

​​ Worker-to-Worker subrequests

To make subrequests from your Worker to another Worker on your account, use Service Bindings. Service bindings allow you to send HTTP requests to another Worker without those requests going over the Internet.

If you attempt to use global fetch() to make a subrequest to another Worker on your account that runs on the same zone, without service bindings, the request will fail.

If you make a subrequest from your Worker to a target Worker that runs on a Custom Domain rather than a route, the request will be allowed.

​​ How many subrequests can I make?

The limit for subrequests a Worker can make is 50 per request on the Bundled usage model or 1,000 per request on the Unbound usage model. Each subrequest in a redirect chain counts against this limit. This means that the number of subrequests a Worker makes could be greater than the number of fetch(request) calls in the Worker.

For subrequests to internal services like Workers KV and Durable Objects, the subrequest limit is 1,000 per request, regardless of the usage model configured for the Worker.

​​ How long can a subrequest take?

There is no set limit on the amount of real time a Worker may use. As long as the client which sent a request remains connected, the Worker may continue processing, making subrequests, and setting timeouts on behalf of that request.

When the client disconnects, all tasks associated with that client’s request are proactively canceled. If the Worker passed a promise to event.waitUntil(), cancellation will be delayed until the promise has completed or until an additional 30 seconds have elapsed, whichever happens first.


​​ Simultaneous open connections

While handling a request, each Worker is allowed to have up to six connections open simultaneously. The connections opened by the following API calls all count toward this limit:

Once a Worker has six connections open, it can still attempt to open additional connections.

  • These attempts are put in a pending queue — the connections will not be initiated until one of the currently open connections has closed.
  • Earlier connections can delay later ones, if a Worker tries to make many simultaneous subrequests, its later subrequests may appear to take longer to start.

If you have cases in your application that use fetch() but that do not require consuming the response body, you can avoid the unread response body from consuming a concurrent connection by using response.body.cancel().

For example, if you want to check whether the HTTP response code is successful (2xx) before consuming the body, you should explicitly cancel the pending response body:

let resp = await fetch(url)
// Only read the response body for successful responses
if (resp.statusCode <= 299) {
// Call resp.json(), resp.text() or otherwise process the body
} else {
// Explicitly cancel it
resp.body.cancel()
}

This will free up an open connection.

If the system detects that a Worker is deadlocked on open connections — for example, if the Worker has pending connection attempts but has no in-progress reads or writes on the connections that it already has open — then the least-recently-used open connection will be canceled to unblock the Worker.

If the Worker later attempts to use a canceled connection, an exception will be thrown. These exceptions should rarely occur in practice, though, since it is uncommon for a Worker to open a connection that it does not have an immediate use for.


​​ Environment variables

The maximum number of environment variables (secret and text combined) for a Worker is 128 variables on the Workers Paid plan, and 64 variables on the Workers Free plan. There is no limit to the number of environment variables per account.

Each environment variable has a size limitation of 5 KB.


​​ Worker size

A Worker can be up to 10 MB in size after compression on the Workers Paid plan, and up to 1 MB on the Workers Free plan.

You can assess the size of your Worker bundle after compression by performing a dry-run with wrangler and reviewing the final compressed (gzip) size output by wrangler:

$ wrangler deploy --outdir bundled/ --dry-run
# Output will resemble the below:
Total Upload: 259.61 KiB / gzip: 47.23 KiB

Note that larger Worker bundles can impact the start-up time of the Worker, as the Worker needs to be loaded into memory. You should consider removing unnecessary dependencies and/or using Workers KV, a D1 database or R2 to store configuration files, static assets and binary data instead of attempting to bundle them within your Worker code.


​​ Worker startup time

A Worker must be able to be parsed and execute its global scope (top-level code outside of any handlers) within 400 ms. Worker size can impact startup because there is more code to parse and evaluate. Avoiding expensive code in the global scope can keep startup efficient as well.


​​ Number of Workers

Unless otherwise negotiated as a part of an enterprise level contract, all paid Workers accounts are limited to a maximum of 500 Workers at any given time. Free Workers accounts are limited to a maximum of 100 Workers at any given time.


​​ Number of routes per zone

Each zone has a limit of 1,000 routes. If you require more than 1,000 routes on your zone, consider using Workers for Platforms or request an increase to this limit.

​​ Number of routed zones per Worker

When configuring routing, the maximum number of zones that can be referenced by a Worker is 1,000. If you require more than 1,000 zones on your Worker, consider using Workers for Platforms or request an increase to this limit.


​​ Image Resizing with Workers

When using Image Resizing with Workers, refer to Image Resizing documentation for more information on the applied limits.


​​ Log size

You can emit a maximum of 128 KB of data (across console.log() statements, exceptions, request metadata and headers) to the console for a single request. After you exceed this limit, further context associated with the request will not be recorded in logs, appear when tailing logs of your Worker, or within a Tail Worker.

Refer to the Workers Trace Event Logpush documentation for information on the maximum size of fields sent to logpush destinations.

Review other developer platform resource limits.