GraphQL requests are cached by default with some GraphQL servers so if it's caching a GraphQL POST t
GraphQL requests are cached by default with some GraphQL servers so if it's caching a GraphQL POST then that's probably why.
I(as in, not at the CDN level, it would be at the Worker/GraphQL library level)
Swhat's worker level? The local server?
II assume you have a Worker serving GraphQL requests there?
Snope, the back end is our graphql server, it does not have any caching implemented for worker
Syes
IWhat GraphQL server is it?
IFor example, Apollo server sometimes has caching enabled by default.
IIt's definitely not the CDN caching it unless you have specifically told the CDN to cache it somehow.
Swe are making some thing like apollo server
SIIt looks like you have a cache of some sort there?
ISyes that's different thing
Sit's entity cache
Syou can use
@cachewith specified ttl
SI thought you were referring to https://github.com/tailcallhq/tailcall/blob/main/cloudflare/src/cache.rs
but anyways even that code is not used, it was recently introduced, the screenshot is quite old
but anyways even that code is not used, it was recently introduced, the screenshot is quite old
GThe comments i get are for another worker that is running, the one under fev is silent/
DIs it possible to force smart placement to be on?
GWhat is that?
Gno one is particualrly talking to anyone
AHey all, question. From the behavior I'm seeing, smart placement does not count requests to R2 as "subrequests", and therefore is not getting past the
Is this expected behavior?
INSUFFICIENT_SUBREQUESTSIs this expected behavior?
CWhat do you mean by force? If you set it then it should be on. You can’t force a location though
SI'm working on a small project i.e.., A user can install there subdomain path form main domain
ex :- hashnode.com //main domain
sunilkumar.hashnode.dev //sub domain like that
we have personal access token to access particular details of a user and later we need to check like what are available domains on the user and we need to take which domain you need to install like that
by using workers how can i achieve this problem .
can anyone help me out the best way to achieve and also some insights and resources
@andrew | as1003
ex :- hashnode.com //main domain
sunilkumar.hashnode.dev //sub domain like that
we have personal access token to access particular details of a user and later we need to check like what are available domains on the user and we need to take which domain you need to install like that
by using workers how can i achieve this problem .
can anyone help me out the best way to achieve and also some insights and resources
@andrew | as1003
MHi.
So, I have used the default workers example for caching POST request: https://developers.cloudflare.com/workers/examples/cache-post-request
The scenario is the following:
So, I have used the default workers example for caching POST request: https://developers.cloudflare.com/workers/examples/cache-post-request
The scenario is the following:
- User lands on the page and the POST request to the endpoint is made via XHR (AJAX) call. The
isCf-Cache-StatusDYNAMIC
- Same user hits this endpoint with the same payload and CF returns the cached result. The
isCf-Cache-StatusHIT - When another user from a different device hits the same endpoint, CF still hits the origin.
Documentation for Cloudflare Workers, a serverless execution environment that allows you to create entirely new applications or augment existing ones …
MThis may be a stupid question but how do I find out if my worker runs terminate early? I had issues with some code only being executed partly but I see no indications of that in logs
HIs the event you are looking at an HTTP-event, or something else?
MIt's a sveltekit application and more specifically a form submission which is processed serverside on +page.server.svelte.
The +page.server.svelte code is executed in a cloudflare worker
There I do 2 supabase DB requests + 2 other API calls to some email api
The +page.server.svelte code is executed in a cloudflare worker
There I do 2 supabase DB requests + 2 other API calls to some email api
HIf you don’t use waitUntil, then as long as you receive a response, the Worker will have completed its task
MI don't use waitUntil. The thing is I could not reproduce the issue yet when I had logs trailing with wrangler. The only thing I see is that the last API call did not happen. But is there any way to see this in the cloudflare logs to be certain this is actually the issue?
HDoes it happen if you run it locally with wrangler?
HYou could also log the API response to see that it does go through
DI wanted to put the Worker in the US region so I wouldn't get billed for non-US traffic from Google Cloud Run lol
HIf it is GeoIP-based, then it should always be US anyway
HThough, can't you just prevent Cloud Run from booting anywhere other than the US?
MWill try to do, but is it the only way to find out if my execution was terminated early?
HAssuming that the logs don't show an error, then I don't think it would be getting cut off. Might be something else being skipped in the code
HIf the request goes through, then it has completed execution, so maybe there are some conditions for an API call that aren't being fulfilled
DCloud Run is booting only in the US. The thing is if the Worker is running outside of the US, the subrequest to Cloud Run is gonna be outside the US for bandwidth billing purposes, no?
HUh... I'm not actually sure how they determine the "region" it is egressing to, and thus the bill. If it is based on IP geolocation, then all Workers should appear as being from the US
HAs for your original question, if Argo detects that an "origin" is in the US, it will spawn itself there. This does require more than one request per Worker invocation though
II believe Cloud Run uses anycast for their URLs
IAnd then they'd bill based on the DC ingressed at
Dwhatttt. lemme check that!
IIf not then I can't imagine how they bill based on region, GeoIP wouldn't be accurate enough for billing purposes
II know for sure Cloud CDN and Cloud Load Balancer does and I was under the impression that the Cloud Run URL is just GCP-managed LB.
Doh wow you're right
CGitHub
Which Cloudflare product(s) does this pertain to? Wrangler core, Miniflare What version(s) of the tool(s) are you using? 3.19.0 [Wrangler] What version of Node are you using? 20.10.0 What operating...
SHey, so I want to use workers with nodejs package called GameDig. I want to call the worker service, and pass the location where should run. It is possible to give to the worker the region code where to execute?
CWorkers aren't nodejs, they run off the Chrome's Javascript engine V8, kind of like a new tab would, in their own isolates. There's some node compat and CF's own socket stuff, but you would need to majorly alter gamedig to use CF's Socket stuff, and it wouldn't work for anything that depends on UDP/CF only supports TCP. So minecraft would be possible, any steam game/using steam server query protocol wouldn't, and you can't tell the worker where to run. You could create a Durable Object and have it run in a few specific regions (<https://developers.cloudflare.com/durable-objects/reference/data-location/#provide-a-location-hint), but it's only a hint to a wide region and CF doesn't have DOs in South America/Africa atm
You can restrict a Durable Object to a jurisdiction, or provide a location hint.
