Cloudflare Developers•3y ago

> Is CF Down? No. > randomly disconnects How often is this happening? e.g. every second, minute, et

Is CF Down?

No.

randomly disconnects

How often is this happening? e.g. every second, minute, etc.

EErisa > Is CF Down? No. > randomly disconnects How often is this happening? e.g. ever...

anurag•12/10/23, 10:45 AM

I haven't measured it like that. Whenever I start up the worker I randomly get this fetch failed.
Because the local dev is so slow, I write the logic in normal .ts file, run it with bun and once it starts working copy the code to workers.

anurag•12/10/23, 10:46 AM

Says waiting for connection.

kian•12/10/23, 10:49 AM

Does this happen with a simple Worker? Like this

export default {
  fetch() { return new Response("foo") }
}

export default {
  fetch() { return new Response("foo") }
}

Kkian Does this happen with a simple Worker? Like this ``` export default { fetch()...

anurag•12/10/23, 10:50 AM

let me check

Kkian Does this happen with a simple Worker? Like this ``` export default { fetch()...

anurag•12/10/23, 10:52 AM

Yes, after it started I made a request and then it errored out again.

..4li Something cursed? Could you please give me a lead?

Hello, I’m Allie!•12/10/23, 12:13 PM

You would have to manually pass the response back, chunk by chunk. If I may ask though, why do you want this?

kian•12/10/23, 1:17 PM

https://developers.cloudflare.com/fundamentals/reference/http-request-headers/#cf-connecting-ip-in-worker-subrequests

kian•12/10/23, 1:17 PM

In cross-zone subrequests from one Cloudflare customer zone to another Cloudflare customer zone, the CF-Connecting-IP value will be set to the Worker client IP address '2a06:98c0:3600::103' for security reasons.

kian•12/10/23, 1:34 PM

That isn't how you should be blocking Workers anyways.

kian•12/10/23, 1:34 PM

You should use the

cf.worker.upstream_zone

cf.worker.upstream_zone

field - https://developers.cloudflare.com/fundamentals/reference/http-request-headers/#cf-worker

Ttanglu ipv6 is annoying but I don't know how to do that I found workers proxy code from...

Hello, I’m Allie!•12/10/23, 1:52 PM

You can add a Firewall Rule to block all Workers Request from any zone but your own: https://dash.cloudflare.com/?to=/:account/:zone/security/waf/custom-rules/new

Hello, I’m Allie!•12/10/23, 1:52 PM

Something like

(cf.worker.upstream_zone ne "your.doma.in")

(cf.worker.upstream_zone ne "your.doma.in")

Action Block

..4li I have 2 kinds of users, the ones who pay and the ones who use manually. To prev...

Hello, I’m Allie!•12/10/23, 2:07 PM

Couldn't you rate limit them instead?

Original message was deleted

Hello, I’m Allie!•12/10/23, 2:09 PM

Do you have an actual origin? Otherwise, your Worker would be responding to requests for

/chrome.svg

/chrome.svg

with your HTML again

..4li Currently the files are stored on OracleCloud buckets, but that wouldn't make an...

Hello, I’m Allie!•12/10/23, 2:48 PM

If your assets are on R2, then you wouldn’t be paying for bandwidth, if that is what you mean

Kkian Does this happen with a simple Worker? Like this ``` export default { fetch()...

anurag•12/10/23, 6:24 PM

any updates on my issue?

Sir Boniface•12/10/23, 8:27 PM

Hello

andrewp7628•12/10/23, 10:03 PM

I'm seeing Cloudflare Workers IP addresses being outside the range of the stated IP addresses at https://www.cloudflare.com/ips/. Specifically, I'm getting connections from a Cloudflare Worker at:

104.28.153.183

104.28.153.183

, which is outside the range of the closest stated CIDR block on the IPs page of:

104.24.0.0/14

104.24.0.0/14

. Am I on the wrong page for whitelisting Cloudflare Worker IPs?

Cloudflare

IP Ranges | Cloudflare

This page is intended to be the definitive source of Cloudflare’s current IP ranges.

Aandrewp7628 I'm seeing Cloudflare Workers IP addresses being outside the range of the stated...

Cyb3r-Jak3•12/10/23, 10:11 PM

Just curious, what makes you think that is a Cloudflare worker coming from that IP address?

CCyb3r-Jak3 Just curious, what makes you think that is a Cloudflare worker coming from that ...

andrewp7628•12/10/23, 10:16 PM

UFW Logs I'm monitoring on the server I'm connecting to from Cloudflare Workers. Was getting the

Error: proxy request failed, cannot connect to the specified address

Error: proxy request failed, cannot connect to the specified address

. error, and noticed block entries originating from that IP. Whitelisting it resolved the error on "cannot connect" error on Cloudflare's end, so I'm highly confident that's coming from the worker.

Aandrewp7628 I'm seeing Cloudflare Workers IP addresses being outside the range of the stated...

ErisaOP•12/10/23, 10:34 PM

104.24.0.0/14 covers 104.24.0.1 to 104.27.255.254, which does not include 104.28.153.183

I looked up the IP and its apparently used for WARP egress, which doesn't really align with what you saw. Are you able to check it again to make sure and, although it can be spoofed over warp, are you able to log any headers like

cf-connecting-ip

cf-connecting-ip

cf-worker

cf-worker

coming from it?

ErisaOP•12/10/23, 10:35 PM

For what its worth this isnt the first time someones mentioned that range, its adjacent to a warp range so theres a lot of people thinking those IPs are meant to be proxy egress

lukas•12/11/23, 4:59 AM

How long before a worker goes to sleep after the last request?

Llukas How long before a worker goes to sleep after the last request?

Unsmart•12/11/23, 5:03 AM

You can only run background tasks if you use context.waitUntil and those can run for 30 seconds at most if thats what you really want to know.

UUnsmart You can only run background tasks if you use context.waitUntil and those can run...

lukas•12/11/23, 5:19 AM

Gotcha, i was just curious if you can have a local cache (new Map) in the worker and i thought if it stays up for x seconds before going to sleep, then the subsequent requests can use that cache

KKlowner Is there something else in your script performing fetches to external services o...

anurag•12/11/23, 5:22 AM

I tried with this. I was still getting the error.

export default {
  fetch() { return new Response("foo") }
}

export default {
  fetch() { return new Response("foo") }
}

masd4•12/11/23, 7:32 AM

The Durable Objects pricing page says "Once the data is removed, the object will be cleaned up automatically by the system.". When can we expect that to happen? I have durable objects that had no data for a week and they are still visible in the API.

Mmasd4 The Durable Objects pricing page says "Once the data is removed, the object will...

DaniFoldi•12/11/23, 8:01 AM

yeah I don't think they actually get cleaned up, or not very often - I've seen some objects retained for months, but haven't checked for a while now

DDaniFoldi yeah I don't think they actually get cleaned up, or not very often - I've seen s...

masd4•12/11/23, 8:17 AM

hmm interesting, maybe they only start cleaning up when it goes into huge numbers because right now I'm only testing with a few entries

Jan Piotrowski (janpio)•12/11/23, 9:31 AM

The worker size limit is documented as "10MB in size after compression" (for paid accounts): https://developers.cloudflare.com/workers/platform/limits/#worker-size
I know from some other places this refers to gzip compression. Does anyone know which compression level is used?
(Reference https://nodejs.org/api/zlib.html#class-options:~:text=zlib.constants.Z_VERSION_ERROR-,Compression%20levels.,-zlib.constants.Z_NO_COMPRESSION)

JJan Piotrowski (janpio)The worker size limit is documented as "10MB in size after compression" (for pai...

DaniFoldi•12/11/23, 10:07 AM

Interestingly I was never worried enough about the bundle size to check, but here's how wrangler calculates the size shown after deploying: https://github.com/cloudflare/workers-sdk/blob/96faca45733b90ca5b5b2a308c85e2dadeb1c6bf/packages/wrangler/src/deployment-bundle/bundle-reporter.ts#L11-L18

Looks like it's using whatever the defaults are, and I'd assume so is the internal storage

DDaniFoldi Interestingly I was never worried enough about the bundle size to check, but her...

Jan Piotrowski (janpio)•12/11/23, 10:08 AM

Thanks, the wrangler code I also checked - but how certain are you this is also how "server" side calculates and validates?

JJan Piotrowski (janpio)Thanks, the wrangler code I also checked - but how certain are you this is also ...

DaniFoldi•12/11/23, 10:10 AM

I don't recall any issues where wrangler says it's <1mb but fails to upload, and there are no open github issues for this matter either, so statistically I'd say it's probably the case, although I have no insight into the internal mechanism and could be the case that everyone just got lucky with their worker size meowshrug