okay, so my PDF parsing code takes >4 seconds to run, are there any tricks to speed up CPU intensive
okay, so my PDF parsing code takes >4 seconds to run, are there any tricks to speed up CPU intensive code?
Jmultiple threads?
Gis there a way to restart workers environment via cloudflare dashboard?
HNot really, since how do you know which isolate to restart?
Gthen how do i fix these errors?
"outcome": "exception",
"scriptName": "mage-api",
"diagnosticsChannelEvents": [],
"exceptions": [
{
"name": "Error",
"message": "The script will never generate a response.",
"timestamp": 1712091400239
}
],
UI am getting this issue, but haven't found a resolutoin.
J@Leo fun excuse to have a play around with rust 
Jit took me a long time to find a random unmaintained javascript library that used pdfjs to successfully dump tabular data out of my PDF
J@Leo I guess I could get multiple threads by having my worker make multiple calls back to itself?
JIf I could split it by page that would work.. but now I have the wonderful problem of figuring out how to do that
Jand the procedure to split it might be as CPU intensive as extracting the text.. who knows
CIf you were using calls to different Durable Object Instances maybe. Service Bindings would just share the same isolate/thread though, and I would imagine calls out to the same script would just land back on the same isolate, since they just go to the same metal/machine
Jahh.. was thinking that might be the case..
SJust curious about this -> I have a worker and when I traceroute from Delhi NCR region(India) in vultr (just a test), i see that it basically terminates the traffic in Singapore. Is this expected?. I have seen questions in the past that asked similar questions (https://community.cloudflare.com/t/cloudflare-peerings-in-india/15755).
CPretty normal yea, lots of people have reported Indian ISPs routing to Singapore. Routing isn't just about the closest place, Cloudflare controls which routes are available to who (which carrier, ix, etc) and ISP has the final control over where it is routed to. This isn't a Workers thing, but a general cdn thing
Gis there a way to restrict swaggerui dashboard access to only local environment? im using @Cloudflare/itty-router-openapi
SThe surprising thing is if I do this in Bangalore, it terminates in SFO according to geolocation look up.
Cthat's not surpising, geolocation with anycast ips like CF is always wrong
SHow do I know the real location? SFO seems irreasonable with that latency
CThat IP exists in more then one location (anycast) so it's hard to assign a single location to it
SYes, Anycast it is!
CIn general, you can go to
/cdn-cgi/trace path of your website, ex: https://d1-a-ut.makemyparty.live/cdn-cgi/trace, and look at colo={airportcode}, and that airport code is the identifider of the location you hitCOf course, you can't do that with most looking glasses, so you just have to guess by ping time or if traceroute or reverse dns exposes something obvious. Sometimes you'll see one of the end IPs be an Internet Exchange making it pretty obvious
SAnd I would have to do this from within India if I want to really locate that colo (in this particular case)
SIs there a cloudflare utility like thatof Vultr that allows me to trace it locally to a region?
Cfor example:
Cif you google
38.10.0.47, you get a peering db result, it's CF's IP for them on Extreme IX in Bangalore, so it's most likely going to Cf's Banglore coloSgotcha!
Cshould probably go to #off-topic or #general-help since this isn't workers related
Aif i am using durable objects for rate limiting should i use an alarm to delete buckets that are inactive?
DIs there a way I can have my Worker "push" data back so that the client doesn't disconnect?
Dis there a 100,000 ms limit or something on a Worker?
J100s is the default timeline for Cloudflare http things
DI increased the proxy timeout to 6000s; I thought that would apply?
DIs there no easy way to increase that?
JHi all, I'm building an API. I need to find some data from a relatively large file (1~3 GB). The file is encrypted, so I also need a functional algorithm to decode the data from the file. What data storage solution should I choose? Thank you.
RIs there a max body size in Cloudflare Workers? I need to handle files (using FormData for example) up to 2GB
S100MB for the free Cloudflare plan https://developers.cloudflare.com/workers/platform/limits/#request-limits
VHello, is it possible to modify HTML of react site using cloudflare workers with HTMLRewriter?
If my understanding is correct cloudflare workers modify HTTP responses.
But on react site there is not a new HTTP request when page is loaded because react works by changing DOM content.
Any help and explanations would be greatly appreciated!
If my understanding is correct cloudflare workers modify HTTP responses.
But on react site there is not a new HTTP request when page is loaded because react works by changing DOM content.
Any help and explanations would be greatly appreciated!
HHTMLRewriter can only modify HTML that exists Server-Side. Generally, React uses a single div, that is then populated on the client side. So unless you can get React to do some of the work on the server, then no
DDwat, I thought more than 5 subrequest connections would just queue up?
HMaybe you are hitting this?
Dna, I am not doing anywhere near 1000 requests
HHow about 50?
Dnope. I'm doing:
5 subrequests to R2 through the bindings.
5 subrequests to the Cache API
2 subrequests to Backblaze B2
If I await between the R2 put's and the Cache API put's, the error goes away.
5 subrequests to R2 through the bindings.
5 subrequests to the Cache API
2 subrequests to Backblaze B2
If I await between the R2 put's and the Cache API put's, the error goes away.
HMight be this?
If the system detects that a Worker is deadlocked on open connections — for example, if the Worker has pending connection attempts but has no in-progress reads or writes on the connections that it already has open — then the least-recently-used open connection will be canceled to unblock the Worker.Otherwise, not sure what it could be
If the Worker later attempts to use a canceled connection, an exception will be thrown. These exceptions should rarely occur in practice, though, since it is uncommon for a Worker to open a connection that it does not have an immediate use for.
DI wonder if there's an undocumented limit still for internal subrequests
HI don't think so? Pretty sure I've run 20+ concurrently before
Dwouldn't that be an exception, not a "Too many subrequests"?
