I imagine it works fine if you're using Bypass since that just goes to normal zone security
I imagine it works fine if you're using Bypass since that just goes to normal zone security
Doh.. I thought Bypass still counted.
EIf you set the policy to Allow and then include Everyone it won’t work anymore
NAttempting to set up a GitHub action for my Cloudflare worker, however I am struggling to find any examples with the bindings to a D1 database. Am I supposed to commit my wrangler.toml file? I would assume that would not be best practice but I may be wrong?
AYep, wrangler.toml is fine to commit 
AProvided you haven't stored any secrets in there, but things like account ID or D1 database ID aren't considered sensitive
NOkay thanks! Yeah I have other secrets in there, but I think I can move them into the github action environment secrets and add them via the github action
OQuestion: if a worker exceeds its 6 simultaneous network connection count, is it the case that it will be killed with outcome
exceededCpu? We're seeing cron workers that last 500ms getting killed, but some of the logging is pointing to simultaneous network connections as a possible culprit and it's hard to figure out if that would report the same way as high CPU or memory usage (esp given cron workers have 15 minutes to complete).OAhhh.
OHmm okay will take a closer look here.
OYeah, 500ms I was looking at a different thing. It still looks like we're closer to 20-25s than 30s.
OBut actually that's wall clock time, the dashboard is saying 500ms CPU time:
O
OThat seems like the next possible culprit but hard to figure out how we'd get there, this is a pretty lightweight process most of whose work is on the network.
OWill dig a little more though, maybe there's something we're missing.
OThank you!
ADo conditional PutObject requests to R2 count as writes, for billing, if the conditon fails?
OStill interested in the answer to this question though: if we push the past 6 connections limit, do we get a different outcome or is it still reported as
exceededCpu?
Cno, it'd just cancel the oldest and if your code wasn't setup to check for it, you'd script error
CIf 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.https://developers.cloudflare.com/workers/platform/limits/#simultaneous-open-connections
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.
OGot it. Ok. Well I think my hypothetical fix for network connections will also potentially help with memory usage so let's see how it goes haha.
O(It went just fine.)
JHow do I stop my worker from connecting to live services when running locally, and instead wait for me to run a local instance of the service?
Dis the local instance of that other service also another Worker, or something that's just HTTP?
JAnother worker over rpc
JI get a warning from wrangler that it's bound to a live service
Cwhen I enable warp vpn I can't wrangler tail
RFrom there you can select the Respect cache option.
GIf my worker is only used for egressing objects from R2, does it make a functional differnce to use the r2 binding vs the s3 api from r2? We create new buckets daily and want to get rid of us having to reploy the workers to add the new bucket binding
HBindings are generally faster than the API, due to not running over HTTP
HNot to say you can't use the S3 API, just that the difference might be noticeable
Gthanks! Will there ever be something to bind to a group or to all workers?
HDynamic Bindings are something they are working on, but not sure if that specifically applies to R2...
GBut great, thanks for the info!
VHi everyone, I'm trying to make a site using Cf pages & workers. I have a SvelteKit app set up correctly and my demo app is running on pages.
Does the adapter automatically take my
Does the adapter automatically take my
*.server.ts and transform it into cf workers?
H:bibicat:
Bherro all. I have a requirement to geosteer requests to a SIP URI depdending on our cx continent location.
I'm not much of a dev... more like a 50 year old infrastructure script kiddie.
Do you think it's just a matter of altering: https://developers.cloudflare.com/workers/examples/country-code-redirect/
to SIP URIs? Something like,
/
export default {
async fetch(request) {
/**
I'm not much of a dev... more like a 50 year old infrastructure script kiddie.
Do you think it's just a matter of altering: https://developers.cloudflare.com/workers/examples/country-code-redirect/
to SIP URIs? Something like,
/
- GEO SIP redirector
* - - Redirect users based on their location to different SIP URIs
export default {
async fetch(request) {
/**
- A map of the SIP URIs to redirect to based on country
- @param {Object} countryMap*/const countryMap = {US: "houston2.voip.ms",EU: "london1.voip.ms",CA: "houston2.voip.ms"};// Use the cf object to obtain the country of the requestconst country = request.cf.country;if (country != null && country in countryMap) {const sipUri = countryMap[country];// Remove this logging statement from your final output.console.log(`Based on ${country}-based request, your user would go to ${sipUri}.`);return new Response(null, { status: 302, headers: { "Location": `sip:${sipUri}` }});} else {// Default redirection if country is not foundreturn new Response(null, { status: 302, headers: { "Location": "sip:aaa.voip.ms" }});}},};
JAnyone know of a transactional email sending service with a nice serverless compatible api client?
Jthe whole world of api clients is such a mess..
Jdocs say "one or more" queues https://developers.cloudflare.com/queues/reference/configuration/#workers
J@João (grimpa) the name of the queue is in the queue property of the message batch https://developers.cloudflare.com/queues/reference/javascript-apis/#messagebatch
DNAny good resources for learning about system design using the capability security model used by worker RPCs? The Cap'n proto FAQ links out to http://zesty.ca/capmyths/usenix.pdf which I'm working through, but also would like to see some real world examples.
For instance, the RPC announcement gives the example of an auth service that returns a User stub on auth, allowing the caller to interact with the user only if auth succeeds. Let's say we wanted to add a todo list to the app. So, we spin up a few workers:
So a call sequence might be:
Maybe I'm overcomplicating this (a simple todo app certainly doesn't need all this), but I'm curious to see what it looks like to go all-in on the capability model.
For instance, the RPC announcement gives the example of an auth service that returns a User stub on auth, allowing the caller to interact with the user only if auth succeeds. Let's say we wanted to add a todo list to the app. So, we spin up a few workers:
So a call sequence might be:
Maybe I'm overcomplicating this (a simple todo app certainly doesn't need all this), but I'm curious to see what it looks like to go all-in on the capability model.
AQuestion since the docs don’t seem to have this clear, is the ja4 fingerprint included in the ja3_hash bot management variable or is this stored in a different variable?
CThis means that something like this:
Would take very long right?
It would do 6 + 6 + 6 + 6 until it reaches the limit of 50 per request
JDoes in a Cloudflare worker trigger, no longer process all requests? A few months ago, that trigger pattern ran on every request, domain.com/style.css and so on, now it only runs on / ?
KNot quite 6+6+6+6
KOnce one finishes, another one can start
KIt's just 6 concurrent, not batches of 6
