Workers are given certain routes on which they operate. If the route does not include /rewrite.html,

Workers are given certain routes on which they operate. If the route does not include /rewrite.html, then it will not be involved in serving the request

HHello, I’m Allie!Workers are given certain routes on which they operate. If the route does not in...

vakno•1/31/24, 10:07 AM

but why is there a need for /rewrite.html route? By my understanding i only need route /test.html

vakno•1/31/24, 10:08 AM

If i will set both routes for this worker there will be redirect loop

Vvakno I want to achieve changing url from /test.html to /rewrite.html but serving the ...

kian•1/31/24, 10:09 AM

That’s not what you’re doing, you’re redirecting them to the new page

Kkian That’s not what you’re doing, you’re redirecting them to the new page

vakno•1/31/24, 10:11 AM

yes and serving response of trigger route

kian•1/31/24, 10:11 AM

On the new page, the Worker isn’t being hit.

vakno•1/31/24, 10:12 AM

what do i need to change?

Kitodi•1/31/24, 11:33 AM

Does anybody uses TCP Stream in cloudflare workers, when I try to run

read()

read()

without server response it got stuck? #TCP Sockets read() function got stuck

Kitodi•1/31/24, 11:47 AM

Is it considered a bug or can it be prevented if it happens?

Rapid•1/31/24, 11:58 AM

Is there a way to tell cloudflare where I want the worker to work? (certain region instead of close to my region)

Rapid•1/31/24, 11:59 AM

More specifically for Workers Routes to leverage API speed from my servers which reside in n.virginia

Rapid•1/31/24, 12:00 PM

I feel like if I my worker goes from my location to NVirginia instead of NVirginia worker goes directly to nvirginia and then sends me the results, nvirginia worker would be fatster

RRapid More specifically for Workers Routes to leverage API speed from my servers which...

Hello, I’m Allie!OP•1/31/24, 12:01 PM

That sounds like Smart Placement. You aren't able to manually tell it where to spawn the Worker, but it should automatically determine which colo to spawn in to get itself closest to your origin(s)

Rapid•1/31/24, 12:01 PM

By default, Workers and Pages Functions are invoked in a data center closest to where the request was received. If you are running back-end logic in a Worker, it may be more performant to run that Worker closer to your back-end infrastructure rather than the end user. Smart Placement automatically places your workloads in an optimal location that minimizes latency and speeds up your applications.

By default, Workers and Pages Functions are invoked in a data center closest to where the request was received. If you are running back-end logic in a Worker, it may be more performant to run that Worker closer to your back-end infrastructure rather than the end user. Smart Placement automatically places your workloads in an optimal location that minimizes latency and speeds up your applications.

Rapid•1/31/24, 12:02 PM

So the worker is indeed doing some back-end logic in the worker, that's the reason I want it to be spawned close to the server

Rapid•1/31/24, 12:02 PM

Is it possible to see logs of my latest worker runs? to know wher eit was spawned?

RRapid Is it possible to see logs of my latest worker runs? to know wher eit was spawne...

Hello, I’m Allie!OP•1/31/24, 12:06 PM

You could

console.log

console.log

the

request.cf.colo

request.cf.colo

field, then view it in live logs/

wrangler tail

wrangler tail

Rapid•1/31/24, 12:10 PM

I'm seeing real-time logs and I can see it

Rapid•1/31/24, 12:10 PM

so, it does it in France (FRA) as it seems

Rapid•1/31/24, 12:10 PM

and I need it to be around NVirginia, you think I should enable Smart Placements?

kian•1/31/24, 12:11 PM

Smart Placement is automatic and runs it based on where the RTT would be lowest based on your subrequests - if you make enough subrequests to a server there, it'll put it closer to there

Rapid•1/31/24, 12:11 PM

okay checking it now

SSam Shouldn't you also count in cpu time?

dave•1/31/24, 1:56 PM

I'm assuming each subrequest takes less than 30ms of CPU time

dave•1/31/24, 1:56 PM

oh wait

dave•1/31/24, 1:56 PM

it would be clocked as if it was one request

dave•1/31/24, 1:56 PM

...weird.

Chaika•1/31/24, 4:10 PM

When a user sends a request, they get routed to a specific machine in a specific colocation/datacenter in a specific PoP. That machine/metal spins up an isolate for your worker, which has its own memory. Connection keepalive will keep them connected to the same metal for a bit, so you might see it "working" and using your in-memory objects repeatedly, but any other client/new connection would be randomly routed to a different machine
https://developers.cloudflare.com/workers/reference/how-workers-works/

How Workers works · Cloudflare Workers docs

The difference between the Workers runtime versus traditional browsers and Node.js.

Chaika•1/31/24, 4:11 PM

Not even cache is per PoP, it's per colo/datacenter. Some PoPs like ORD are made up of mutiple colos each with their own cache (called multi-colo pops, MCPs)

Chaika•1/31/24, 4:14 PM

but yea the guarantee is really only per request

Isolates are resilient and continuously available for the duration of a request, but in rare instances isolates may be evicted.

Even if they connect to the same metal/machine I believe, they might connect to a different isolate

Because there is no guarantee that any two user requests will be routed to the same or a different instance of your Worker, Cloudflare recommends you do not use or mutate global state.

quotes from: https://developers.cloudflare.com/workers/reference/how-workers-works/

Chaika•1/31/24, 4:14 PM

You can use the Cache API to get per-colo cache
https://developers.cloudflare.com/workers/runtime-apis/cache/

Rapid•1/31/24, 4:16 PM

Guys, what is the expected time to get a response from workers?

I am seeing a weird timing issue.

I have created an API route that return {}. empty object.

When I request it from my location (Europe to NVirginia) I get the response within 180ms.
When a worker requests that data, the worker takes 400 to 490ms to return what took me 180ms. (It uses COLO - FRA)

Anyidea why there's such a big overhead when it comes to it? How can I debug all those times to make sure?

Hhash My plan was to store each index in a dedicated worker. If my understanding is co...

Chaika•1/31/24, 4:18 PM

If you mean hardcoding the "index" (an array of objects, you said?) in the worker code, if its quick enough to load, CF will warm up your worker during the tls connect and in a lot of cases elimate cold starts:
https://blog.cloudflare.com/eliminating-cold-starts-with-cloudflare-workers
If it's anything that can change/calculated off request/cache sort of thing, you'd want Durable Objects most likely

RRapid Guys, what is the expected time to get a response from workers? I am seeing a w...

Rapid•1/31/24, 4:36 PM

Is there a difference between free version and paid version?

RRapid Guys, what is the expected time to get a response from workers? I am seeing a w...

Chaika•1/31/24, 4:38 PM

If your API Proxied?
In a worker you could time the request

const start = performance.now();
const response = await fetch("https://developers.cloudflare.com/");
const end = performance.now();
const timing = end - start; // duration of the subrequest to developers.cloudflare.com

Could you share the endpoints? Curious if I can repro
Might just be a routing thing

Rapid•1/31/24, 4:38 PM

Can I DM you for the endpoint?

Chaika•1/31/24, 4:38 PM

Sure

SSDD•1/31/24, 4:38 PM

I noticed that requests are cached with workers by default, how does that work? Is it at the browser level or with cf?

SSSDD I noticed that requests are cached with workers by default, how does that work? ...

SSDD•1/31/24, 4:39 PM

can I get some links

Hhash If my understanding is correct, There is persistence between requests if loading...

Chaika•1/31/24, 4:40 PM

There's always persistence between requests to the same isolate. The thing is each metal/machine has its own isolates, each colo/cf dc has its own metals, some Pops have mutiple colos, and besides your own connection persistence there's no routing preference to go to a machine that already has your isolate spun up on it

SSSDD can I get some links

SSDD•1/31/24, 4:40 PM

is it https://developers.cloudflare.com/cache/concepts/default-cache-behavior/?

Default Cache Behavior · Cloudflare Cache (CDN) docs

Cloudflare respects the origin web server’s cache headers in the following order unless an Edge Cache TTL cache rule overrides the headers.

CChaika Sure

Rapid•1/31/24, 4:41 PM

Sent all necessary data in DM

By default, Workers and Pages Functions are invoked in a data center closest to where the request was received. If you are running back-end logic in a Worker, it may be more performant to run that Worker closer to your back-end infrastructure rather than the end user. Smart Placement automatically places your workloads in an optimal location that minimizes latency and speeds up your applications.

Workers are given certain routes on which they operate. If the route does not include /rewrite.html,

Similar Threads

Similar Threads

Similar Threads