is it possible to run a cf worker entirely on my infrastructure without incurring charges on cloudfl
is it possible to run a cf worker entirely on my infrastructure without incurring charges on cloudflare?
D*in production
HParts of them, yes. Entirely, no
HYou can run Workers with
wranglerworkerdBut you won't get access to all the different bindings/other goodies that you would on Cloudflare
Dsome cf bindings work with wrangler when it's running on my machine (i assume its calling a cloudflare api), shouldn't it also work on my own infrastructure then?
HYes, but not all of the local emulators are production-ready(nor are they meant to be)
HAnd some of them just flat-out don't exist at all locally
Mrandom question for someone. in a worker, when you call
Response.redirect()
IA response.redirect will simply redirect to whatever you put inside the
redirect()For example, a
Response.redirect("/foo")whats odd is I am, but they are doubling up
ICan you check the actual 301/302 response to see what the Location header is?
Mso I have this redirect request function in my worker and Im logging out what the url is. the url looks correct and says, for example,
my.domain.com?utmtest=123
M
JWhat arguments are you passing to
redirectRequesturlsearch
BHow does cloudflare handling spa website? since it's caching by url. which in my belief it will caching multiple version of the spa application right?
QHow do you choose between durable objects and r2? The docs don't say much about DO
DIn my app I am consuming tasks form a queue taking about half a minute per task. After the task is done, I'd like to notify the user via sse. My current implementation got me so far that I can already have an SSE that notifies the user, but just because the event source route checks against my database every second.
What is the piece I am missing to let the event source worker know when my queue worker finished? How can I connect those?
What is the piece I am missing to let the event source worker know when my queue worker finished? How can I connect those?
AYou can use service bindings to communicate between 2 worker scripts: https://developers.cloudflare.com/workers/configuration/bindings/about-service-bindings/
Service bindings are an API that facilitate Worker-to-Worker communication via explicit bindings defined in your configuration.
DIs that what I want though? If I call the worker normally, I could easily end up on a different location right?
DMeaning the worker that I call from the queue-consumer will not know about the open SSE connection
AYou would probably need to use durable objects too then, so that your callback from the queue script reaches the same instance with the SSE connection
DI'll take a look at durable objects, thanks 
DOkay seems like DO could solve my problem, but is so much complexity necessary just so I can basically get a notice inside the SSE route whenever the queue consumer finishes? It's even the same worker, I wouldn't even need a service binding. But they can easily be on different locations
DBasically I want the EventEmitter to work without caring that there is in fact an infrastructure with different isolates. Does anyone know how I can have similar API without too much back and forth that I'd have with DO?
QHow do you choose between durable objects and r2? The docs don't say much about DO.
JHello! I am trying to change the Content-Length header, as I will use Cloudflare Workers to download files from a R2 bucket. I tried by changing the header manually, but it doesnt make a change. Then I read that it was not possible to change that header, and that the only way to do so is by using Streams.
Now, my question is how should I do that? I tried to do it (see the code below), however it does not work. Any advice is very much appreciated!
Now, my question is how should I do that? I tried to do it (see the code below), however it does not work. Any advice is very much appreciated!
QThis is the doc that I specifically didn't find useful. The DO section didn't really tell me when I'd use it over R2. I feel like you need to already know what the company had in mind for DO before you use it. Like maybe it's some kind of popular storage design that I'm not familiar with, so I don't know what it's for.
DI am using a worker for server-sent events. Basicall it reads from a database every 3 seconds and sends the results via SSE to the client. Now if I understand correclty, the new standard pricing will neither evict quickly (like the old bundled plan) nor will it be high duration costs (because only cpu time is important) right?
CR2 is better for bigger files as it has the largest storage size. DO is good when you need all the data to be synced across the global.
AEviction can happen on any plan but usually doesn’t happen unless there’s a Cloudflare update (about once a day in my experience), free plan users get evicted more frequently than paid/Enterprise. And yes it’s charged in CPU milliseconds instead of wall time.
DThank you
.For a worker route, if you don’t setup any rate limiting
.What’s like the, default rate limit set to
KFree is 1000 reqs per min
KPaid doesn't have one (that limits the whole Worker)
QThose don't seem related though. 
What if I need small files (~2kB) that serve more as debug logs and will rarely-to-never be accessed?
What if I need small files (~2kB) that serve more as debug logs and will rarely-to-never be accessed?HI would still go for R2. DOs are more for when you need compute attached to storage
KDurable Objects are stateful Workers that have transactional storage attached
KR2 is just object storage, strongly consistent with an S3 API
KBy far the simplest to use, since the S3 API means you can interact with it outside of Workers/the CF REST API
QI'm only using Workers to write the data, but I don't intend to read much from it.
QBut "compatible with S3" doesn't really mean much if I'm not using S3.
KMeans you can back it up easily, interact with it outside of Workers (i.e your own device), migrate to other storage options easily, etc
KIt's not about you using S3, it's about how many things integrate or work with the S3 API
QDon't both R2 and DO advertise as strongly consistent? If that's not a tie-breaker between them, what is?
QI'm not currently using, nor going to use, anything that integrates with the S3 API, so this doesn't help me choose between them. 
QThe limit page is a bit better than the differences page. 
Looks like R2 is for just dumping data into it [free tier + unlimited storage]. DO says it's a paid-only feature that limits you to 50GB for your entire account.
Looks like R2 is for just dumping data into it [free tier + unlimited storage]. DO says it's a paid-only feature that limits you to 50GB for your entire account.
