CCorrect. You need a paid zone to be able to open tickets. Everyone is able open Account, billing and registrar tickets.
WWhat are you seeing?
SIf I am using Workers Websocket without Durable Objects (for example using it for listening to queues or as a proxy for another websocket server), how would the pricing work? Is it charged based on per request (hence per websocket connection) and the cpu-time used to process the websocket messages?
FBoth CPU time and per Worker request (not subrequests the worker does)
SMax of 30 seconds of CPU time per invocation - is this happening on the websocket ?
IYes. But keep in mind this is CPU time, not wall time (not time as measured on a clock but time that is measured based on how much compute you are doing)
II also believe a Websocket message resets that limit? (someone would have to confirm though)
FWebsockets seem to have indeed a very long or infinite timeout if I read this correctly -> https://community.cloudflare.com/t/how-can-i-use-a-worker-for-real-time-stuff-limits/325031/2?u=fry69
SOkay, I tested it, it stays longer than 30 seconds. a 2-minute connection was provided.
I30 seconds is cpu time.
ILet me see, I believe I had a tag for this.
I?cpu-time
FCPU Time: the amount of work the CPU needs to do, measured in
Wall Time: the actual amount of time a Worker runs for, as though you measured it with a clock or your browser DevTools
msfetchWall Time: the actual amount of time a Worker runs for, as though you measured it with a clock or your browser DevTools
FUnless you have an endless for loop running in your worker to eat CPU time, you might have just measured real (wall) time, which does not matter. Workers can run for a year I think if they do not consume CPU, see -> https://developers.cloudflare.com/workers/platform/limits/#duration
IYou get 30 seconds of cpu time, so that means that if your WebSocket isn't doing much it could theoretically last for many minutes, although it's not a 100% because sometimes websockets are killed by Cloudflare for internal reasons.
S"wall time" is not included in the fee ?
Sand execution duration ?
FCorrect. Wall time does not matter, only CPU time.
For billing only CPU and number of Worker requests matter (not subrequests).
For billing only CPU and number of Worker requests matter (not subrequests).
Sthere is no pricing based on durable object logic, is there?
Sis there any link explaining?
FSI'm confused about this.
SCan this method be similar to the durable object pricing?
Sbecause 1 request fee is 1 hour connection and more
Sam I wrong?
1I kinda assumed that WebSockets on Workers would just get charged at the same rate as WebSockets on DOs, by GB-sec
Uwasnt sure where to post this so i assume this is where, but, is it possible to entirely limit access to a cloudflared tunnel to a worker?
Ui couldnt find a zero trust channel but i also have terrible eyesight
Unevermind i just made a WAF for it
Ywhat's the best way to do this with reverse proxy
I want to take url /academy?id=25
show content of this page on
/academy/academy-name
25 will change to name using json
I want to take url /academy?id=25
show content of this page on
/academy/academy-name
25 will change to name using json
0Cloudflare Docs
Set multiple Cron Triggers on three different schedules.
0I think cloudflare is extremly under priority on documentation
SI agree. It can be priced that way. unfortunately, there is no complete explanation.
0anyone knows how do i access to d1 database from cron trigger?
0hey i got it, cause im using hono
0abit confused
0is anyone using cron trigger, does it work for you lol
00by testing it, am i going to see
console.log("cron processed");cron processedyup
0i did exactly 100% same as the tutorial
0let me try with your sample code
0i saw someone discussed this in github
0.
0i created another file cron.ts, and pasted the same code, then i realised it was all wrong
0even if i am using workers, this poor documentation doesnt specify where you should store your code
0i would create a new file anyways by jumping into this documentation
0
0which part mentioned it should be merged into the
index.tswhy cant just admit poor documentation was maintained
