Does anyone have any idea of how much workers for platforms cost and what kind of limits they have?
Does anyone have any idea of how much workers for platforms cost and what kind of limits they have?
H- Pricing: Contact Sales
- Limits
CThanks!
MIll take a peek. I added Lerna and that seemed to work fine for our use case, but Im interested to solve the cross-worker dependencies since I think the docs are outdated for that. if I come across solutions maybe Ill open a pr or something
KYou'd have to deploy the Worker everytime, but that's probably the best way to do it anyways.
KFetching R2/KV everytime vs updating the Worker once a week
KThat's still a deployment
KYou can use it though, yeah.
KWhat’s the issue with doing a deployment though?
KIf it's small enough to fit into your Worker without any issues, including it in the deployment is faster & cheaper
KWell, if you need it on every request.
JCan I do something like this using itty-router in Workers?
KJOh thank you
JThats what I was looking for
DUsing a non-R2 S3 API is oddly easier to cache, since you just have to use cacheEverything: true on the fetch.
DI use Backblaze B2 (with the S3 API) for some Workers that use caching heavily.
DHello there,
So I am using the workers to merge images together which takes about 1ms and then to upload it to the R2 storage (with the put command) which (at least wall time) takes about 500ms.
My app may make 5, 10 or 100 post requests without waiting for the previous one to finish and that's fine by me because it takes a lot of time to upload the image to R2.
The issue is, because of that, after just 5 or 10 requests I start to get 'service temporarily unavailable.'. Am I supposed to wait for the worker to finish, can I not make concurrent requests?
So I am using the workers to merge images together which takes about 1ms and then to upload it to the R2 storage (with the put command) which (at least wall time) takes about 500ms.
My app may make 5, 10 or 100 post requests without waiting for the previous one to finish and that's fine by me because it takes a lot of time to upload the image to R2.
The issue is, because of that, after just 5 or 10 requests I start to get 'service temporarily unavailable.'. Am I supposed to wait for the worker to finish, can I not make concurrent requests?
Dps: please tag me on the reply
AYou can check if they were caught by the rate limit from the Events section in Security on the Cloudflare Dashboard. I believe it’s 1,000 requests per minute
DI headed over to my website and then security and analytics and this is what I see
D
HTry adding some logging to the Worker, then tailing it. Is it the Image Merging that is erroring? Or the put to R2? Or something else?
DSure, I can add multiple logs and check it out and get back to you on that. Something that pinged my curiosity, can I edit the limits for my ip (or my server's for that matter), I wouldn't want cf to limit the number of requests on my prod server.
DBased on what Ace mentioned, I thought there weren't any.
HThe 1k requests that they are talking about are the API Ratelimits(afaict), which wouldn't apply here, since you are using the Workers APIs. They should be able to handle substantially higher rates of throughput
HThe API Ratelimits only apply to requests hitting
api.cloudflare.comDAh alright, so that shouldn't apply to me since I have connected my worker to a custom domain.
DI will use the wrangler to tail my worker and add a few extra errors to let you know.
DI don't think the image merging is erroring though.
HIt's more that you are(hopefully) using the R2 bindings in Workers, which aren't subject to the same(if any) rate-limiting
HThat should apply even if you use
workers.devDI am using the bindings, alright, I'll tag you when I have a few logs.
Thi guys, has anyone tried yugabyte db with cf workers? Is its performance better than D1 or PlanetScale? Since D1 does not handle table altering and typing really well, D1 also does not have local backup (although the alpha backend had it) and can't natively integrate with external service so I'm considering switching to MySQL and Postgres
DSo, if I make the requests on their own they don't exceed the cpu time, which is the error thrown (which is set at 5kms for this testing). When I am making the requests in batches of 20 or more though, then the requests towards the end start failing. They should have their own 5kms limit though as they are new requests which they don't even come close to.
Below are some examples from various logs:
Same images, just another request
Below are some examples from various logs:
Same images, just another request
DIs there a limit somewhere that I need to update?
UI don’t think those specific limits will be changed tbh
SHey, is there a recommended way to protect workers with api keys? I'm fairly new to API gateways as a whole and saw that cloudflare offers an API gateway through shield, but it's only available to enterprise clients
Was wondering if there was a way to setup a separate worker to act as a proxy and for handling api key management while storing the api keys in KV storage
Was wondering if there was a way to setup a separate worker to act as a proxy and for handling api key management while storing the api keys in KV storage
SAlso came across this post on reddit that talked about how to implement just that, but it goes really in depth
https://www.reddit.com/r/CloudFlare/comments/194r1kx/cloudflare_worker_review_api_key_service/
https://www.reddit.com/r/CloudFlare/comments/194r1kx/cloudflare_worker_review_api_key_service/
Reddit
Explore this post and more from the CloudFlare community
