Vercel's Edge Functions are built on top of Cloudflare Workers. The runtime is designed for any similar provider, but we've chosen Cloudflare for now because they have an amazing product.

Hacker News

•

10/26/21, 5:41 PM

HHello, I’m Allie!https://news.ycombinator.com/item?id=29003514

✦

✦ xcaeser•9/19/23, 3:09 PM

thx for sharing, reading now

✦

✦ xcaeser•9/19/23, 3:10 PM

okay let me try the worker with vercel kv in this case and see if it's better

✦

✦ xcaeser•9/19/23, 3:29 PM

how to get around this? Uncaught ReferenceError: process is not defined
cloudflare workers don't use node right?

✦

✦ xcaeser•9/19/23, 3:34 PM

nvm, followed upstash cloudfare quickstart

HHello, I’m Allie!Also, wouldn't this more be a comparison of CF KV vs. Upstash?

✦

✦ xcaeser•9/19/23, 3:37 PM

you're right, switched to upstach. Cloudflare KV is very very slow

✦

✦ xcaeser•9/19/23, 3:41 PM

still aw seems to be faster for now;

Unsmart•9/19/23, 3:50 PM

you show 2 responses for the other platform one of which is faster one of which is slower so

✦✦ xcaeser still aw seems to be faster for now;

Chaika•9/19/23, 3:52 PM

What are you measuring? Puts or Gets?

should've gone for foundation db

The magic with KV is its unlimited scale and getting faster the more requests you give it, it's made for read-heavy cases, like for example Cloudflare Pages uses it
https://developers.cloudflare.com/kv/learning/how-kv-works/

You can signal for it to cache results longer as well, and stream back responses, ex

   let value = await env.KV.get("index.html", {cacheTtl: 86400, type: "stream" });
    return new Response(value, {
      headers: {
        "content-type": "text/html"
      }
    })

   let value = await env.KV.get("index.html", {cacheTtl: 86400, type: "stream" });
    return new Response(value, {
      headers: {
        "content-type": "text/html"
      }
    })

(It is eventually consistent by design, but by getting it to cache for longer you're not going to see updates until cache expires or if you write in same location. (KV 2.0 which was delayed aimed to solve this by checking every minute in the background for updates, but it was delayed due to issues))
It's pretty fast, once its cached locally ~20ms or less

CChaika What are you measuring? Puts or Gets? > should've gone for foundation db The mag...

✦

✦ xcaeser•9/19/23, 5:00 PM

understood thx chaika

dave•9/20/23, 5:16 AM

What are your preferred ways of securely accessing a Worker route from an AWS Lambda? I’ve done the reverse before, undecided on what makes the most sense for the other way around.

Ddave What are your preferred ways of securely accessing a Worker route from an AWS La...

sathoro•9/20/23, 7:33 AM

good old secret key in the headers

Ddave What are your preferred ways of securely accessing a Worker route from an AWS La...

Hello, I’m Allie!OP•9/20/23, 7:36 AM

mTLS?

Amnon•9/20/23, 8:07 AM

Hey,
I'm using Workers to push some logic to the request and proxying the request to the origin backend
When combining that with the Dynamic URL Rewrite Transform Rule it seems like the Transform Rule is triggered two times - from the client to the worker and when proxying the request
In my Dynamic Transform Rule, I used - concat(“/v1”, http.request.uri.path)concat(“/v1”, http.request.uri.path) and ended up with /v1/v1/<original_req_path>/v1/v1/<original_req_path>

I tried to adjust the Custom Filter Expression with no success (skipping the rule every time the URL starts with /v1/v1 works but doesn't feel robust enough)
Any idea how to overcome this behavior?
Thanks

(also posted to the community - https://community.cloudflare.com/t/proxyrequest-in-workers-triggers-transform-rules-more-than-once/557777)

javeed•9/20/23, 9:02 AM

Hi folks,

I have worker route in my site example (miton.com)::

route value us wildcard / and it is pointed to some worker

Here all the requested from any domain or wildcard domain connected with this site are coming to worker route

I need to by pass some of domain example "javi.milton.com" not to go to worker route

Because i have pointed "javi.militon.com" to R2 bucket

javeed•9/20/23, 9:03 AM

I have create new route for "javi.milton.com" and disabled route withour choosing any workers

javeed•9/20/23, 9:03 AM

but still no luck, it is keep pointing to / --> workers

kian•9/20/23, 9:48 AM

You can use Unbound, which gives you up to 30 seconds of CPU time.

Original message was deleted

DaniFoldi•9/20/23, 9:54 AM

if you only do hashing on the client, you're essentially storing the "raw" password, just transformed, which I don't recommend

DaniFoldi•9/20/23, 9:56 AM

all hashing algorithms transform the password, and the theory behind it is interesting, but it's different

DaniFoldi•9/20/23, 9:57 AM

a malicious actor that knows the clienthash can get access, vs if you hash on the server, you still can't do anything if you know the hash

Proprietary storage engine afaik

Similar Threads

Similar Threads

Similar Threads