```js export class MyContainer extends Container { defaultPort = 8080; sleepAfter = this.env.SLE

Kkian ```js export class MyContainer extends Container { defaultPort = 8080; sleep...

N

Nicolasmp•8/26/25, 5:59 PM

Alright, and if that variable is undefined it will default to never sleeping?

K

kianOP•8/26/25, 6:00 PM

The default value is 10 minutes so that

S

Sean•8/26/25, 7:09 PM

Wondering if anyone has some guidance they can share - we have an AI process that runs in a container and takes about 15-20 minutes per invocation. This is triggered by an R2 event subscription where the consumer worker calls my Cloudflare Container app. Right now the issue is the proxy worker that forwards the request to the container is timing out at 180,000 ms (3 minutes) and causing the process to be retried. The actual process works within the deployed container, but we don't want it to be continuously retried by the platform timeout error. This seems to be the worker CPU time limit. Unfortunately pushing it up to 5 minutes won't help.

SSean Wondering if anyone has some guidance they can share - we have an AI process tha...

C

Chaika•8/26/25, 7:21 PM

Cpu time is not consumed by waiting on i/o like a fetch. If the client stays connected, there's no limit on duration. Do they all timeout at exactly 3 minutes? That doesn't really match any platform limit that comes to mind, hmm. I assume you're using a queue and calling the container worker via rpc from the queue consumer?

CChaika Cpu time is not consumed by waiting on i/o like a fetch. If the client stays con...

S

Sean•8/26/25, 7:32 PM

Yes, correct. I have a message worker (queue consumer) that makes a request to my CF Container app. The worker associated with the container app consistently errors out at ~180k ms

SSean Yes, correct. I have a message worker (queue consumer) that makes a request to m...

C

Chaika•8/26/25, 7:35 PM

Queue Consumers should have a max of 15 minutes duration time. Proxy Worker should stay alive as long as the request is pending. The container will sleep if no requests in 10 minutes by default, unless you've configured sleepAfter to 3 minutes.
If you click on the event, is there any more info on the error?

CChaika Queue Consumers should have a max of 15 minutes duration time. Proxy Worker shou...

S

Sean•8/26/25, 7:37 PM

Here's the full error

SSean Here's the full error

C

Chaika•8/26/25, 7:38 PM

That's from the container worker, not the queue consumer

CChaika That's from the container worker, not the queue consumer

S

Sean•8/26/25, 7:39 PM

Correct, that's where I'm getting the errors (the container worker). Apologies if I misstated.

SSean Correct, that's where I'm getting the errors (the container worker). Apologies i...

C

Chaika•8/26/25, 7:45 PM

Your queue consumer should be awaiting the fetch/rpc and then returning success, otherwise it'd just be building up the queue over and over. Are you doing something weird like waitUntil in it and ignoring errors?

S

Sean•8/26/25, 7:53 PM

@Chaika This is my queue consumer code:

Error details attempting to hit the container app start-async endpoint:

Msg 2000d8094472de33684f1967e7365521: Error processing message (attempt 1) for d1cab0d6-be82-4211-a6a9-c53c00dbff40: Error: HTTP error! status: 500 - {"error":"Internal Server Error"}

Msg 2000d8094472de33684f1967e7365521: Error processing message (attempt 1) for d1cab0d6-be82-4211-a6a9-c53c00dbff40: Error: HTTP error! status: 500 - {"error":"Internal Server Error"}

Msg 2000d8094472de33684f1967e7365521: Error processing message (attempt 1) for d1cab0d6-be82-4211-a6a9-c53c00dbff40: Error: HTTP error! status: 500 - {"error":"Internal Server Error"}

Msg 2000d8094472de33684f1967e7365521: Error processing message (attempt 1) for d1cab0d6-be82-4211-a6a9-c53c00dbff40: Error: HTTP error! status: 500 - {"error":"Internal Server Error"}

SSean @Chaika This is my queue consumer code: Error details attempting to hit the co...

C

Chaika•8/26/25, 8:18 PM

hmm, yea I don't think this is a container thing. Your worker log says cancelled, meaning the calling worker gave up on it.
Using a layer inbetween/Mastra Client is going to complicate getting the underlying error out, a quick google search shows there's a configurable timeout for Mastra which a default, worth looking into

CChaika hmm, yea I don't think this is a container thing. Your worker log says cancelled...

S

Sean•8/26/25, 8:19 PM

Ok thanks for your help

D

Dipan•8/27/25, 1:31 AM

Hey, I am deploying a Server in cloudflare container, deployed it successfully and get this response -

Total Upload: 53.47 KiB / gzip: 13.58 KiB
Worker Startup Time: 14 ms
Your worker has access to the following bindings:
- Durable Objects:
  - MORPH_CONTAINER: MorphContainer
- Vars:
  - NODE_ENV: "production"
Uploaded morph-server (9.65 sec)
Deployed morph-server triggers (1.76 sec)

Total Upload: 53.47 KiB / gzip: 13.58 KiB
Worker Startup Time: 14 ms
Your worker has access to the following bindings:
- Durable Objects:
  - MORPH_CONTAINER: MorphContainer
- Vars:
  - NODE_ENV: "production"
Uploaded morph-server (9.65 sec)
Deployed morph-server triggers (1.76 sec)

Total Upload: 53.47 KiB / gzip: 13.58 KiB
Worker Startup Time: 14 ms
Your worker has access to the following bindings:
- Durable Objects:
  - MORPH_CONTAINER: MorphContainer
- Vars:
  - NODE_ENV: "production"
Uploaded morph-server (9.65 sec)
Deployed morph-server triggers (1.76 sec)

Total Upload: 53.47 KiB / gzip: 13.58 KiB
Worker Startup Time: 14 ms
Your worker has access to the following bindings:
- Durable Objects:
  - MORPH_CONTAINER: MorphContainer
- Vars:
  - NODE_ENV: "production"
Uploaded morph-server (9.65 sec)
Deployed morph-server triggers (1.76 sec)

But when I am seng a request to the server getting this error (attached in screenshot)

D

Dipan•8/27/25, 1:31 AM

Can anyone help me here?

DDipan Hey, I am deploying a Server in cloudflare container, deployed it successfully a...

I

Isaac McFadyen•8/27/25, 1:33 AM

How are you calling the container? Sounds like your Worker code is throwing the error before the request makes it to the container.

D

Dipan•8/27/25, 1:34 AM

This is the workser/index.js file

message.txt2.04KB

D

Dipan•8/27/25, 1:36 AM

worker/index.js*

IIsaac McFadyen How are you calling the container? Sounds like your Worker code is throwing the ...

D

Dipan•8/27/25, 1:39 AM

Can you send me any resources, Where I can see that how can I make a ideal worker?

A

A Shy Cow•8/27/25, 10:36 AM

hi everyone

is there any containers cost calculator? i need a container which is up 24/7 and i'd like to know how much i would pay for it

A

Alexander•8/27/25, 12:31 PM

I am trying to run a dev server and point wrangler to pull an image from cloudflare registry but run into an error - anyone know how to solve?

∷ Building container images for local development...
                                                                                                                                                                                                                                                  
error when starting dev server:
TypeError: Failed to parse URL from /registries/registry.cloudflare.com/credentials                                                                                                                                                                
    at node:internal/deps/undici/undici:13510:13                                                                                                                                                                                                   
    at async sendRequest (file:///home/azuev/polymer/node_modules/.pnpm/@cloudflare+vite-plugin@1.12.0_vite@7.1.3_@types+node@24.3.0_jiti@2.5.1_lightningcss@1._8b43f7dcedb891a211abe2294734dcbb/node_modules/@cloudflare/vite-plugin/dist/index.js:14260:10)                                                                                                                                                                                                                                         
    at async file:///home/azuev/polymer/node_modules/.pnpm/@cloudflare+vite-plugin@1.12.0_vite@7.1.3_@types+node@24.3.0_jiti@2.5.1_lightningcss@1._8b43f7dcedb891a211abe2294734dcbb/node_modules/@cloudflare/vite-plugin/dist/index.js:14320:26    
 ELIFECYCLE  Command failed with exit code 1.
➜  polymer git:(main) ✗

∷ Building container images for local development...
                                                                                                                                                                                                                                                  
error when starting dev server:
TypeError: Failed to parse URL from /registries/registry.cloudflare.com/credentials                                                                                                                                                                
    at node:internal/deps/undici/undici:13510:13                                                                                                                                                                                                   
    at async sendRequest (file:///home/azuev/polymer/node_modules/.pnpm/@cloudflare+vite-plugin@1.12.0_vite@7.1.3_@types+node@24.3.0_jiti@2.5.1_lightningcss@1._8b43f7dcedb891a211abe2294734dcbb/node_modules/@cloudflare/vite-plugin/dist/index.js:14260:10)                                                                                                                                                                                                                                         
    at async file:///home/azuev/polymer/node_modules/.pnpm/@cloudflare+vite-plugin@1.12.0_vite@7.1.3_@types+node@24.3.0_jiti@2.5.1_lightningcss@1._8b43f7dcedb891a211abe2294734dcbb/node_modules/@cloudflare/vite-plugin/dist/index.js:14320:26    
 ELIFECYCLE  Command failed with exit code 1.
➜  polymer git:(main) ✗

∷ Building container images for local development...
                                                                                                                                                                                                                                                  
error when starting dev server:
TypeError: Failed to parse URL from /registries/registry.cloudflare.com/credentials                                                                                                                                                                
    at node:internal/deps/undici/undici:13510:13                                                                                                                                                                                                   
    at async sendRequest (file:///home/azuev/polymer/node_modules/.pnpm/@cloudflare+vite-plugin@1.12.0_vite@7.1.3_@types+node@24.3.0_jiti@2.5.1_lightningcss@1._8b43f7dcedb891a211abe2294734dcbb/node_modules/@cloudflare/vite-plugin/dist/index.js:14260:10)                                                                                                                                                                                                                                         
    at async file:///home/azuev/polymer/node_modules/.pnpm/@cloudflare+vite-plugin@1.12.0_vite@7.1.3_@types+node@24.3.0_jiti@2.5.1_lightningcss@1._8b43f7dcedb891a211abe2294734dcbb/node_modules/@cloudflare/vite-plugin/dist/index.js:14320:26    
 ELIFECYCLE  Command failed with exit code 1.
➜  polymer git:(main) ✗

∷ Building container images for local development...
                                                                                                                                                                                                                                                  
error when starting dev server:
TypeError: Failed to parse URL from /registries/registry.cloudflare.com/credentials                                                                                                                                                                
    at node:internal/deps/undici/undici:13510:13                                                                                                                                                                                                   
    at async sendRequest (file:///home/azuev/polymer/node_modules/.pnpm/@cloudflare+vite-plugin@1.12.0_vite@7.1.3_@types+node@24.3.0_jiti@2.5.1_lightningcss@1._8b43f7dcedb891a211abe2294734dcbb/node_modules/@cloudflare/vite-plugin/dist/index.js:14260:10)                                                                                                                                                                                                                                         
    at async file:///home/azuev/polymer/node_modules/.pnpm/@cloudflare+vite-plugin@1.12.0_vite@7.1.3_@types+node@24.3.0_jiti@2.5.1_lightningcss@1._8b43f7dcedb891a211abe2294734dcbb/node_modules/@cloudflare/vite-plugin/dist/index.js:14320:26    
 ELIFECYCLE  Command failed with exit code 1.
➜  polymer git:(main) ✗

Curiously I have successfully built and pushed an image - so I am not sure what is the issue?

A

Alexander•8/27/25, 1:26 PM

Another thing I've noticed (hesitant to call it an issue) but calling .getState() returns stale values if:
- in development I manually stop the docker container instance OR ctrl +c the dev server
- impact: when I try to call / fetch container next time, I check via .getState() which returns a stale value (whatever the state was before instance was manually stopped)

I would have thought that .getState() always returns a fresh state but without booting the container up / waiting for ports. Am I wrong?

I previously stored state flag (i.e. running / healthy / error) in DO KV storage using lifecycle hooks but ran into the same stale values issue.

Am I missing something?

SSean Wondering if anyone has some guidance they can share - we have an AI process tha...

S

sampotter•8/28/25, 2:47 AM

I've had this issue for a while. In my limited understanding from testing it looks like container lifecycles are tied to durable objects. If the durable object is evicted from memory, the container is killed also. I have never been able to keep a durable object alive for more than 180,000ms. It looks like that is some sort of hard limit on DOs.

S

sampotter•8/28/25, 2:48 AM

Started a thread here that never got anywhere containers-beta

S

sampotter•8/28/25, 2:50 AM

Our solution in the interim is to make the container workload durable to restarts (which doesn't sound like it's possible for your workload). We also run a loop within the durable object to automatically awake it after it is evicted.

M

mike•8/28/25, 9:53 PM

are other folks still hitting this as well? it continues to be a pain point for us deploying changes - containers-beta

Mmike are other folks still hitting this as well? it continues to be a pain point for ...

K

kianOP•8/28/25, 9:59 PM

Is

(request id: undefined):  Internal Server Error

(request id: undefined):  Internal Server Error

the exact error you're getting?

Kkian Is `(request id: undefined): Internal Server Error` the exact error you're gett...

M

mike•8/28/25, 10:02 PM

yes, just after the edit to update the image:

Error rolling out application <app> due to an internal error (request id: undefined):   Internal Server Error

Error rolling out application <app> due to an internal error (request id: undefined):   Internal Server Error

K

kianOP•8/28/25, 10:06 PM

What's the application ID (UUID in the dashboard URL when viewing a Containers app)?

K

kianOP•8/28/25, 10:07 PM

It might also be somewhere else in the Wrangler output but I'm not sure

Kkian What's the application ID (UUID in the dashboard URL when viewing a Containers a...

M

mike•8/28/25, 10:08 PM

a03fcf42-2bed-4ce5-92d6-cdecb864ac94

a03fcf42-2bed-4ce5-92d6-cdecb864ac94

Mmike `a03fcf42-2bed-4ce5-92d6-cdecb864ac94`

K

kianOP•8/28/25, 10:47 PM

Thanks, I'll ask the team to take a look. Do you have anything in your Wrangler configuration around rollout percentages?

Kkian Thanks, I'll ask the team to take a look. Do you have anything in your Wrangler ...

M

mike•8/28/25, 10:51 PM

we do not, this service is a bit unusual in that we only run a single instance that is triggered via

scheduled

scheduled

on a cron interval (in case that is relevant). thanks!

K

kianOP•8/28/25, 10:55 PM

looks like the team has a fix for that error on the way

I

Ian•8/29/25, 9:59 PM

Just wanna give a feedback that it's kind of worrying that a network egress fee is being slapped on, Cloudflare is known for zero egress fees to begin with..

M

max•8/30/25, 5:49 PM

Is it possible to get containers running in vitest? I can get most things hooked up, but

ctx.container

ctx.container

ctx.container

ctx.container is always undefined. Thanks!

J

jrdn•8/30/25, 7:18 PM

Is there a way to have a container run like a sidecar next to my worker so it’s always served from the same machine for lowest latency possible?

Jjrdn Is there a way to have a container run like a sidecar next to my worker so it’s ...

H

Hello, I’m Allie!•8/30/25, 7:53 PM

No. Not all edge servers are capable of hosting containers. You may get lucky with placement, but it can’t be guaranteed

V

vin•9/1/25, 12:59 PM

Hi, I modified my code to call container with a certain ID "constant-id-234", and modifed the max instances from 15 to 1. Ideally, I think all containers will sleep except "constant-id-234". But there is an container with ID "79dfd71efb78d86b9fcf0293eaac276e49008ad34f84bb2f1e52f9be170e33d6" is still keeping running now. And I cannot access the container via the new version code, an exception was throwed "Failed to start container: The maximum amount of containers allowed to run in the Durable Object namespace is the same to the amount of running Durable Object Containers, we won't be able to assign an instance to this Durable Object Container". Has anyone encountered a similar situation?

Kkian It'll come from the IPs on <https://www.cloudflare.com/en-gb/ips/> but otherwise...

B

berkant•9/1/25, 3:55 PM

Would this possibly have any implications for origins with firewall set up to allow inbound from these IPs only, considering the fact with Containers you're able to egress from the same set of IPs to arbitrary destinations?

R

Ray Li•9/2/25, 5:56 PM

hi! forwarding my previous question here in case someone has done this:

pulumi worker script w/ container binding

J

Jonathan•9/3/25, 5:03 PM

I'm attempting to understand the networking behavior for containers. Currently my container exposes a zero trust tunnel with a custom dns name tied to a cloudflare zone. Attempting to access the webserver running within the container is running very very slow (feels equivalent to dial up internet).

I've attempted to determine if it was just when accessing it via the cloudflare tunnel, however I'm noticing similar behavior when I fetch through the durable object to the container.

Is there an optimization or setting I need to get reasonable networking speed?

JJonathan I'm attempting to understand the networking behavior for containers. Currently m...

K

kianOP•9/3/25, 5:22 PM

There aren't really any related settings, the bandwidth should be pretty good for most apps.

Are you seeing slow cold starts or the actual HTTP response being streamed slowly?

Kkian There aren't really any related settings, the bandwidth should be pretty good fo...

J

Jonathan•9/3/25, 6:55 PM

Specifically the HTTP response is slow for both cases of: performing a fetch to the container (I have a default port specified and it hits a basic python fast api which returns information about the environment) as well as hitting the container using the custom domain to return a web app (verified the web app is not the cause, this is simply just a very slow response as if it takes a long time to go from the zero trust tunnel back to the container origin server) then a delay before the bits come back, the actual processing within the container is pretty quick

P

pandalambda2•9/4/25, 10:41 AM

Why isn't the container FAQ growing? Support, discord, plus all the other channels looks like there's enough uncertainity that perhaps should be updated every other day, so much to cover, so many questions unanswered..

P

pandalambda2•9/4/25, 10:44 AM

On rollouts https://developers.cloudflare.com/containers/faq/#how-do-container-updates-and-rollouts-work if I have max_instances = 1 in my wrangler json config, and my container id is stable (i.e. just 'default

Cloudflare Docs

Frequently Asked Questions

Frequently Asked Questions:

P

pandalambda2•9/4/25, 10:44 AM

does this guarantee the container behaves like a singleton, during rollouts it doesnt try to bring up a replacement first?

S

Steps•9/4/25, 2:32 PM

Aside from https://developers.cloudflare.com/containers/platform-details/#limits , do Containers have any other limits (like invocations per month, etc.) or are these bound to the calling worker? So could I direct all 10 M workers calls to a container without extra costs or is there something to be aware of?

S

Steps•9/4/25, 3:01 PM

For example, if I set the timeout to 10 minutes... will one call then cause that much execution time removed from the worker?

```js export class MyContainer extends Container { defaultPort = 8080; sleepAfter = this.env.SLE

Similar Threads

Similar Threads

Similar Threads