a0325cd9-6679-4d75-b1fe-4bea193f14ec

Kkian Depending what it is, I'd probably say opening an issue on <https://github.com/c...

J

jmickelly•8/20/25, 6:11 PM

Hello, I contacted earlier about whether if Cloudflare Containers supports nested virtualization in their Container infrasturcture and need the functionality for allowing using microVMs inside of the container.

Is there a way of customizing how the container itself exposes devices such as /dev/kvm?

TTobiLG a0325cd9-6679-4d75-b1fe-4bea193f14ec

K

kian•8/20/25, 7:56 PM

I've passed it onto the team. I don't suppose you have anything else like a timestamp or the image tag of the deployments that took a long time?

G

gruntlord6•8/20/25, 11:12 PM

is there any way to kill a container that is stuck running?

G

gruntlord6•8/20/25, 11:13 PM

is deleting it the only option if the process controlling it became detached?

P

Prydar•8/21/25, 12:23 AM

have they said anything about raising container limits for the beta? they're kinda low for our testing/purposes

C

China•8/21/25, 5:40 PM

Hi, I’m running into an issue with my Worker + Containers setup. In my wrangler.toml, I’ve defined both the container and the durable object binding like this:

[[env.production.containers]]
max_instances = 10
class_name = "ForumFullscanContainer"
image = "./containers/forum-fullscan/ForumFullscanDockerfile.dockerfile"

[[env.production.durable_objects.bindings]]
name = "FORUM_FULLSCAN_CONTAINER"
class_name = "ForumFullscanContainer"

[[env.production.containers]]
max_instances = 10
class_name = "ForumFullscanContainer"
image = "./containers/forum-fullscan/ForumFullscanDockerfile.dockerfile"

[[env.production.durable_objects.bindings]]
name = "FORUM_FULLSCAN_CONTAINER"
class_name = "ForumFullscanContainer"

[[env.production.containers]]
max_instances = 10
class_name = "ForumFullscanContainer"
image = "./containers/forum-fullscan/ForumFullscanDockerfile.dockerfile"

[[env.production.durable_objects.bindings]]
name = "FORUM_FULLSCAN_CONTAINER"
class_name = "ForumFullscanContainer"

[[env.production.containers]]
max_instances = 10
class_name = "ForumFullscanContainer"
image = "./containers/forum-fullscan/ForumFullscanDockerfile.dockerfile"

[[env.production.durable_objects.bindings]]
name = "FORUM_FULLSCAN_CONTAINER"
class_name = "ForumFullscanContainer"

I deployed locally yesterday using npx wrangler deploy --env=production, and it was working fine. Today, after another deployment via our GitHub Action, it stopped working. The wrangler.toml hasn’t changed. and the Durable Object migration has also been applied.
The error I’m seeing in the Worker logs is:
"Container is not enabled for this durable object class. Have you correctly setup your wrangler.toml?"
Could someone help me understand why this might be happening?

J

joeyblake•8/22/25, 12:40 AM

Hey y'all! I have been trying to get a very simple worker and container running locally, and when the container is created, it gets a really high

HOST_PORT

HOST_PORT

HOST_PORT

HOST_PORT I'm using rancher desktop. And my docker file builds correctly and is usable on its own. But when running

wrangler dev

wrangler dev

then hitting the endpoint I see the container start up in rancher, but the

HOST_PORT

HOST_PORT

HOST_PORT

HOST_PORT is not what I would expect it to be. I'm assuming that's why I can't connect when using the local worker. I really appreciate any insight y'all may have.

G

goofysystem•8/22/25, 12:42 AM

Running into this on deploy:

--- 2025-08-22T00:33:52.056Z debug
Checking for local image registry.cloudflare.com/<account-id>/rust-task-container-staging:90ba2969 failed with error: Expected there to only be 1 valid digests for this repository: registry.cloudflare.com/<account-id>/rust-task-container-staging but there were 0
---

--- 2025-08-22T00:34:28.883Z log

---

--- 2025-08-22T00:34:28.934Z error
[31m✘ [41;31m[[41;97mERROR[41;31m][0m [1mError rolling out application rust-task-container-staging due to an internal error (request id: undefined):   Internal Server Error[0m


---

--- 2025-08-22T00:34:28.935Z debug
Error: Error rolling out application rust-task-container-staging due to an internal error (request id: undefined):   Internal Server Error
    at apply (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:64300:17)
    at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
    at async deployContainers (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:64435:5)
    at async deploy (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:152882:5)
    at async Object.handler (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:80948:66)
    at async Object.handler (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:77880:7)
---

--- 2025-08-22T00:33:52.056Z debug
Checking for local image registry.cloudflare.com/<account-id>/rust-task-container-staging:90ba2969 failed with error: Expected there to only be 1 valid digests for this repository: registry.cloudflare.com/<account-id>/rust-task-container-staging but there were 0
---

--- 2025-08-22T00:34:28.883Z log

---

--- 2025-08-22T00:34:28.934Z error
[31m✘ [41;31m[[41;97mERROR[41;31m][0m [1mError rolling out application rust-task-container-staging due to an internal error (request id: undefined):   Internal Server Error[0m


---

--- 2025-08-22T00:34:28.935Z debug
Error: Error rolling out application rust-task-container-staging due to an internal error (request id: undefined):   Internal Server Error
    at apply (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:64300:17)
    at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
    at async deployContainers (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:64435:5)
    at async deploy (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:152882:5)
    at async Object.handler (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:80948:66)
    at async Object.handler (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:77880:7)
---

--- 2025-08-22T00:33:52.056Z debug
Checking for local image registry.cloudflare.com/<account-id>/rust-task-container-staging:90ba2969 failed with error: Expected there to only be 1 valid digests for this repository: registry.cloudflare.com/<account-id>/rust-task-container-staging but there were 0
---

--- 2025-08-22T00:34:28.883Z log

---

--- 2025-08-22T00:34:28.934Z error
[31m✘ [41;31m[[41;97mERROR[41;31m][0m [1mError rolling out application rust-task-container-staging due to an internal error (request id: undefined):   Internal Server Error[0m


---

--- 2025-08-22T00:34:28.935Z debug
Error: Error rolling out application rust-task-container-staging due to an internal error (request id: undefined):   Internal Server Error
    at apply (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:64300:17)
    at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
    at async deployContainers (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:64435:5)
    at async deploy (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:152882:5)
    at async Object.handler (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:80948:66)
    at async Object.handler (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:77880:7)
---

--- 2025-08-22T00:33:52.056Z debug
Checking for local image registry.cloudflare.com/<account-id>/rust-task-container-staging:90ba2969 failed with error: Expected there to only be 1 valid digests for this repository: registry.cloudflare.com/<account-id>/rust-task-container-staging but there were 0
---

--- 2025-08-22T00:34:28.883Z log

---

--- 2025-08-22T00:34:28.934Z error
[31m✘ [41;31m[[41;97mERROR[41;31m][0m [1mError rolling out application rust-task-container-staging due to an internal error (request id: undefined):   Internal Server Error[0m


---

--- 2025-08-22T00:34:28.935Z debug
Error: Error rolling out application rust-task-container-staging due to an internal error (request id: undefined):   Internal Server Error
    at apply (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:64300:17)
    at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
    at async deployContainers (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:64435:5)
    at async deploy (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:152882:5)
    at async Object.handler (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:80948:66)
    at async Object.handler (/Users/myname/projects/work/utc/apps/edge-service/node_modules/wrangler/wrangler-dist/cli.js:77880:7)
---

rust-task-container-staging already exists and can be viewed from the dashboard btw.

EDIT: Ok found a way to reproduce this. wrangler deploy will always fail the very first time when there is no image and generate an image tagged registry.cloudflare.com. If I keep the image and run deploy, it will succeed with

Image already exists remotely, skipping push

Image already exists remotely, skipping push

Image already exists remotely, skipping push

Image already exists remotely, skipping push. If you delete the image from Docker Desktop and run deploy you will run into the same error. This explains why it always fails in ci.

M

Maurice•8/22/25, 7:10 AM

When working inside containers there is no access to worker bindings, so when using a product like queues, limited to pushing by sdk / http. This works, however, when working on development (wrangler dev), the simulated queues are in the run time and not exposed to http, so the only option is to debug in production. Is there a better way to get around this? I'm not sure this is the best place to report this, but I figured someone would pick it up.

S

scotto•8/22/25, 1:54 PM

it would be great to have a simple doc on deploing a single dockerfile, similar to what flyio offsets, like Dockerfile -> deploy, simple, current docs are mostly focues on the workers + containers integrations

Ggoofysystem Running into this on deploy: ``` --- 2025-08-22T00:33:52.056Z debug Checking fo...

G

gruntlord6•8/22/25, 8:27 PM

I always have to deploy twice lol

MMaurice When working inside containers there is no access to worker bindings, so when us...

G

gruntlord6•8/22/25, 8:28 PM

Can't you just have the durable object make those requests?

Ggruntlord6 I always have to deploy twice lol

G

goofysystem•8/22/25, 8:31 PM

So is everyone deploying locally bc ci builds the image afresh everytime so it will always fail

. but also it used to work before so

G

gruntlord6•8/22/25, 8:32 PM

no it seems to work fine either though CI/CID when I push with github. If using wrangler it works the second time I push

G

gruntlord6•8/22/25, 8:33 PM

I have noticed a higher error rate from wrangler the last few days, both on my machine and the CI/CID pipeline. Most of them internal errors or "malformed responses"

Ggruntlord6 no it seems to work fine either though CI/CID when I push with github. If using...

G

goofysystem•8/22/25, 8:36 PM

I use wrangler in ci too thats probably why. I really want to stay using wrangler to deploy from ci so hopefully fix comes soon

Ggoofysystem Running into this on deploy: ``` --- 2025-08-22T00:33:52.056Z debug Checking fo...

G

goofysystem•8/23/25, 3:07 AM

Found out that wrangler deploy does not replace the image with the new image ever. The image shown on cloudflare containers dashboard was much older version regardless of how many new image I deploy. Obviously, the first failure uploads to the repository but fails to update the image to use the newst tag. And the second run just skips pushing the new image and treats it as already existing. Only way to get new image up is deleting the container from the dashboard and redeploying, which is not ideal.

B

bentlegen•8/23/25, 3:51 AM

Came in here to basically repeat the same as above.

"✘ [ERROR] Error rolling out application <appname> due to an internal error (request id: undefined): Internal Server Error"

We basically can't auto-deploy new containers anymore from CI. Pretty painful.

B

bentlegen•8/23/25, 4:09 AM

My latest attempt was to just try deploying 2x ... no luck.

T

TobiLGOP•8/23/25, 9:41 AM

I see a lot of RuntimeStartFailed messages when attempting to deploy as well... Deployments take >15min now for me

G

gruntlord6•8/23/25, 3:40 PM

I see a lot of errors about not being able to allocate containers, then they allocate anyway apprently

TTobiLG I see a lot of RuntimeStartFailed messages when attempting to deploy as well... ...

G

gruntlord6•8/23/25, 3:40 PM

as in the rolling deployments?

G

gruntlord6•8/23/25, 3:41 PM

or the actual push to cloudflare

Ggruntlord6 as in the rolling deployments?

T

TobiLGOP•8/23/25, 4:27 PM

Pre-existing application with updated images…

TTobiLG Pre-existing application with updated images…

G

gruntlord6•8/23/25, 4:47 PM

They take a while for me as well, and they are not exactly large images

I

iddogino•8/24/25, 1:21 AM

Experiencing a new issue and wondering if anyone has experienced it too. I have worker A conencting to worker B using a service binding (with

fetch

fetch

), and worker B uses a container. It used to work (last tested about a week ago), but as of yesterday it doesn't seem to work at all. Worker A makes the fetch request and it just gets stuck.
1. Running both locally with wrangler - everything works as expected (I can see A fetching from B locally).
2. Making a regular request to B works (when I switch to

fetch("https://worker-b.account.workers.dev/process/video"

fetch("https://worker-b.account.workers.dev/process/video"

fetch("https://worker-b.account.workers.dev/process/video"

fetch("https://worker-b.account.workers.dev/process/video" it works fine).

Anyone seeing something similar?

Iiddogino Experiencing a new issue and wondering if anyone has experienced it too. I have ...

D

darkpool•8/24/25, 2:46 AM

does

wrangler tail

wrangler tail

wrangler tail

wrangler tail show any logs?

Ddarkpool does `wrangler tail` show any logs?

I

iddogino•8/24/25, 4:09 PM

I can see the request going out from A but no request makes it to B...

T

TobiLGOP•8/25/25, 2:15 PM

Has anyone else problems when redeploying container projects? E.g. I see the following errors:

Checking for local image flexstore:6601d7c5 failed with error: Cannot read properties of undefined (reading 'split')

Checking for local image flexstore:6601d7c5 failed with error: Cannot read properties of undefined (reading 'split')

Checking for local image flexstore:6601d7c5 failed with error: Cannot read properties of undefined (reading 'split')

Checking for local image flexstore:6601d7c5 failed with error: Cannot read properties of undefined (reading 'split'). If I delete the containers app in the Dashboard, I can deploy again

T

Tory•8/25/25, 4:26 PM

What would cause container instances to just stop. These instances have been running steadily, without any code changes for several weeks. Then suddenly they all completely died.

D

deved9036•8/25/25, 7:47 PM

are there issues with deploying containers? just stopped working all of a sudden today for me

D

deved9036•8/25/25, 7:48 PM

seems to be mainly containers, also the observability is throwing a bunch of errors all of a sudden

T

TobiLGOP•8/25/25, 9:34 PM

yeah, same for me. Have you tried updating to the latest wrangler?

D

deved9036•8/25/25, 9:41 PM

i'll give it a god now

D

deved9036•8/25/25, 9:51 PM

okaayy nerverming! seems that fixed it lmao, works on 4.32.0

D

deved9036•8/25/25, 9:52 PM

@TobiLG you still having problem on the latest?

Ddeved9036 @TobiLG you still having problem on the latest?

T

TobiLGOP•8/25/25, 9:54 PM

not currently, but it was pretty unstable the last few days

D

deved9036•8/25/25, 9:55 PM

yeah, for sure!

J

julian7797•8/25/25, 11:33 PM

When using a container from another durable object, I appear to be running into issues, but from inside regular worker, all is well? Anyone else seen this

D

danamajid•8/26/25, 10:03 AM

hi @Mike Nomitch | Workers PM i'm getting:
"The image registry does not exist"

according to https://www.answeroverflow.com/m/1389977616509047009
you shared it's because the account was part of the private beta (which i my case it was).

can you pls help?

Anyone seen this before when pushing - Cloudflare Developers

Anyone seen this before when pushing their image?

npx wrangler@latest containers images list

 ⛅️ wrangler 4.22.0
───────────────────

Error listing images: ApiError: The image registry does not exist

npx wrangler@latest containers images list

 ⛅️ wrangler 4.22.0
───────────────────

Error listing images: ApiError: The image registry does not exist

npx wrangler@latest containers images list

 ⛅️ wrangler 4.22.0
───────────────────

Error listing images: ApiError: The image registry does not exist

npx wrangler@latest containers images list

 ⛅️ wrangler 4.22.0
───────────────────

Error listing images: ApiError: The image registry does not exist

supposed to be using the built-in registry, maybe I've missed something obvious

N

Nicolasmp•8/26/25, 11:57 AM

Hi there! Is there any way to have different

sleepAfter

sleepAfter

sleepAfter

sleepAfter variables according to env variables? For example, I want my container to stop after an hour in dev but never shutdown in prod.

NNicolasmp Hi there! Is there any way to have different `sleepAfter` variables according to...

K

kian•8/26/25, 5:58 PM

I think you can just set it to a

this.env.???

this.env.???

this.env.???

this.env.??? value

K

kian•8/26/25, 5:58 PM

export class MyContainer extends Container {
  defaultPort = 8080;
  sleepAfter = this.env.SLEEP_AFTER;
}

export class MyContainer extends Container {
  defaultPort = 8080;
  sleepAfter = this.env.SLEEP_AFTER;
}

export class MyContainer extends Container {
  defaultPort = 8080;
  sleepAfter = this.env.SLEEP_AFTER;
}

export class MyContainer extends Container {
  defaultPort = 8080;
  sleepAfter = this.env.SLEEP_AFTER;
}

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

kian•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?

a0325cd9-6679-4d75-b1fe-4bea193f14ec

Similar Threads