Hey all! I tried to find sandbox instance limits but there is no clear answer in the docs anywhere -

AAlexander Hey all! I tried to find sandbox instance limits but there is no clear answer in...

G

gruntlord6•8/17/25, 6:00 PM

sanbox?

G

gruntlord6•8/17/25, 6:00 PM

do you mean the max instances you define?

A

AlexanderOP•8/17/25, 8:36 PM

yep referring to @Cloudflare/sandboxes

aren't max instances referring to max container instances? or 1 sanbox == 1 container instance used?

K

Khagan•8/17/25, 9:35 PM

Hello, I'm using Next.js + Django. Is it possible to deploy this on cloudflare containers?

AAlexander yep referring to @Cloudflare/sandboxes aren't max instances referring to max c...

1

1984 Ford Laser•8/17/25, 10:14 PM

Yeah a sandbox is a container instance

1

1984 Ford Laser•8/18/25, 7:49 AM

Any news on containers allowing arbitrary TCP/UDP traffic? RTMP in/out is a major use case for me

T

TobiLG•8/18/25, 12:55 PM

I currently see very slow container deployments, around 10min from what I can see in th Containers Dashboard... And then it takes a few minutes afterwards until I no longer see

A call to blockConcurrencyWhile() in a Durable Object waited for too long. The call was canceled and the Durable Object was reset

A call to blockConcurrencyWhile() in a Durable Object waited for too long. The call was canceled and the Durable Object was reset

A call to blockConcurrencyWhile() in a Durable Object waited for too long. The call was canceled and the Durable Object was reset

A call to blockConcurrencyWhile() in a Durable Object waited for too long. The call was canceled and the Durable Object was reset

errors. Is this to be expected currently? Edit: It took nearly 15min after the intial deployment until requests were served

M

Marco Aleixo•8/18/25, 2:22 PM

hey team - is there a limit of subrequests in containers as well?

D

darkpool•8/18/25, 7:18 PM

do containers support grpc? Couldnt find any mention of it in docs. If not, is support on roadmap?

M

MrBithles•8/18/25, 7:30 PM

Is the concept containers only for local/isolated tasks that can be executed by a worker.
As I've been struggling to get a simple container to run which can handle requests to call external sites with fetch requests or Playwright.
Issues like:

Singletons objects in worker are not scoped to code that runs in container, its obv to say but not obv in code what is scoped.
Contains can't do network requests? (
```
✘ [ERROR] Error: [unenv] https.request is not implemented yet!
```
✘ [ERROR] Error: [unenv] https.request is not implemented yet!
```
✘ [ERROR] Error: [unenv] https.request is not implemented yet!
```
✘ [ERROR] Error: [unenv] https.request is not implemented yet!
)
Not sure why examples/templates as for hono app vs the worker request class, I was initially treating index.ts as the entry to Docker

The way of getting a container seems odd, following the example of using getRandom always throws an error first:

✘ [ERROR] Error checking if container is ready: connect(): Connection refused: container port not found. Make sure you exposed the port in your container definition.
✘ [ERROR] Error checking if container is ready: The operation was aborted
Container exited with unexpected exit code: 1 {"exitCode":1,"name":"Error","stack":"Error: Container exited with unexpected exit code: 1"}
[wrangler:info] POST / 500 Internal Server Error (2153ms)
✘ [ERROR] Error: [unenv] https.request is not implemented yet!

✘ [ERROR] Error checking if container is ready: connect(): Connection refused: container port not found. Make sure you exposed the port in your container definition.
✘ [ERROR] Error checking if container is ready: The operation was aborted
Container exited with unexpected exit code: 1 {"exitCode":1,"name":"Error","stack":"Error: Container exited with unexpected exit code: 1"}
[wrangler:info] POST / 500 Internal Server Error (2153ms)
✘ [ERROR] Error: [unenv] https.request is not implemented yet!

✘ [ERROR] Error checking if container is ready: connect(): Connection refused: container port not found. Make sure you exposed the port in your container definition.
✘ [ERROR] Error checking if container is ready: The operation was aborted
Container exited with unexpected exit code: 1 {"exitCode":1,"name":"Error","stack":"Error: Container exited with unexpected exit code: 1"}
[wrangler:info] POST / 500 Internal Server Error (2153ms)
✘ [ERROR] Error: [unenv] https.request is not implemented yet!

✘ [ERROR] Error checking if container is ready: connect(): Connection refused: container port not found. Make sure you exposed the port in your container definition.
✘ [ERROR] Error checking if container is ready: The operation was aborted
Container exited with unexpected exit code: 1 {"exitCode":1,"name":"Error","stack":"Error: Container exited with unexpected exit code: 1"}
[wrangler:info] POST / 500 Internal Server Error (2153ms)
✘ [ERROR] Error: [unenv] https.request is not implemented yet!

T

TobiLG•8/18/25, 10:12 PM

Deployments really seem to be oscilliating today.... Earlier the containers were available within a minute again, but now >10min and nothing... How can this be debugged? Is this expected?

MMrBithles Is the concept containers only for local/isolated tasks that can be executed by ...

K

kian•8/19/25, 12:06 AM

That unenv error sounds more related to your Worker than the Container itself

TTobiLG Deployments really seem to be oscilliating today.... Earlier the containers were...

K

kian•8/19/25, 12:21 AM

What exactly are you measuring, the time when making a request to an inactive instance?

Kkian What exactly are you measuring, the time when making a request to an inactive in...

T

TobiLG•8/19/25, 5:22 AM

Recurrent curl requests to the endpoints, until the service gets available. Had everything between 1-20mins yesterday. And that is after the Dashboard says the deployments are done… Honestly, this is not really tolerable for a Container Service when people want to iterate quickly during development. Also, sometimes it seems that I receive cached responses, even though all the old instances are shown to be shut down, which is strange. I‘m aware that this is in beta, but multi-minute downtimes upon redeployments might be a showstopper.

T

TobiLG•8/19/25, 5:26 AM

I‘ve been part of the private beta, and I didn’t see this behavior before. Unsure if there‘s something I can do to improve this on my end, or how to debug this. The Worker logs show the timeouts to DOs, but no error messages

N

Nick•8/19/25, 9:50 AM

Does anyone know what happens if I redeploy the DO while a container is running? Will the container be killed?

M

Meathill•8/19/25, 3:37 PM

Hello, everyone. I tried to setup a Strapi Docker on worker, but now I am stuck at

Failed to start container: The container is not listening in the TCP address 10.0.0.1:1337

Failed to start container: The container is not listening in the TCP address 10.0.0.1:1337

Failed to start container: The container is not listening in the TCP address 10.0.0.1:1337

Failed to start container: The container is not listening in the TCP address 10.0.0.1:1337
Anyone can help me? Thank you all

Jjschlesser make sure you have a line like this in your Dockerfile `EXPOSE 1337` and dont fo...

M

Meathill•8/19/25, 4:22 PM

Yes, I think so, I have test this Image locally and it works well

Jjschlesser locally use localhost instead of 10.0.0.1 , you will have to do some environment...

M

Meathill•8/19/25, 4:24 PM

I tried 0.0.0.0 before, and it didn't work. So

localhost

localhost

should be better?

Jjschlesser what is the code you are using to connect?

M

Meathill•8/19/25, 4:25 PM

Sorry, I don't understand. What code do you mean? I hope to setup a Strapi docker as my backend, currently I cannot visit from browser.

M

Meathill•8/19/25, 4:32 PM

OK, I have used

docker build

docker build

locally and

docker run

docker run

locally to make sure my

Dockerfile

Dockerfile

works.
Then I followed the tutorial from Cloudfalre to create this worker.
The entry is quite simple, like a few lines of code:

NNick Does anyone know what happens if I redeploy the DO while a container is running?...

K

kian•8/19/25, 4:52 PM

I suspect it's handled the same as any other rollout: https://developers.cloudflare.com/containers/faq/#how-do-container-updates-and-rollouts-work

TTobiLG I‘ve been part of the private beta, and I didn’t see this behavior before. Unsur...

K

kian•8/19/25, 4:53 PM

What's the application ID (the GUID in the URL when viewing the instances page for that Containers app)?

I can pass it onto the team & ask them if they see anything unusual.

K

kian•8/19/25, 5:07 PM

Depending what it is, I'd probably say opening an issue on https://github.com/cloudflare/containers/issues as the best way to get it to the team

G

gruntlord6•8/19/25, 8:22 PM

This was an interesting one, had a degraded instance. Its gone now but it keeps occupying a slot of capacity even when all the instances are done

G

gruntlord6•8/19/25, 8:25 PM

its finally let go of it now but that was bizzare

Ggruntlord6 This was an interesting one, had a degraded instance. Its gone now but it keeps ...

K

kian•8/20/25, 12:44 AM

The team said it’s just a UI bug (they don’t count as live) which they’ll fix

Kkian The team said it’s just a UI bug (they don’t count as live) which they’ll fix

G

gruntlord6•8/20/25, 1:23 AM

Weird, it wouldn’t spin a 10th instance as far as I could tell

Jjschlesser i just tried to connect from my DO to 10.0.0.1 in my local environment and it wo...

M

Meathill•8/20/25, 2:30 AM

I think your suggestion worked somewhat. I got another error now:

Failed to start container: The container is not running, consider calling start()

Failed to start container: The container is not running, consider calling start()

Failed to start container: The container is not running, consider calling start()

Failed to start container: The container is not running, consider calling start()
I think maybe it's because that the Strapi need some time to get fully started, so the request failed

MMeathill I think your suggestion worked somewhat. I got another error now: `Failed to sta...

T

TobiLG•8/20/25, 5:25 AM

Had the same yesterday. It Took several minutes until the container was available

TTobiLG Had the same yesterday. It Took several minutes until the container was availabl...

M

Meathill•8/20/25, 5:47 AM

Hey, thanks, so I should refresh the page until the container got fully started and the webpage shows?

Kkian What's the application ID (the GUID in the URL when viewing the instances page f...

T

TobiLG•8/20/25, 12:15 PM

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.

Hey all! I tried to find sandbox instance limits but there is no clear answer in the docs anywhere -

Similar Threads

Similar Threads

Similar Threads