Thanks, they are sure fast starting scraping, like instant, hope they do not eat up too much of the
Thanks, they are sure fast starting scraping, like instant, hope they do not eat up too much of the free tier
CIt can help to disable the workers.dev subdomain for a bit when first creating the worker. Means the bots will get rejected when they first try it
Nfor storing profile pictures and other kinds of pictures (like product images uploaded), should I use R2 or Cloudflare Images (stored)?
Nwhich is more worth it?
AThanksss for the tip !
SHey everyone, MailChannels will be discontinuing our free API service as of June 30th. Details are in this knowledgebase article: https://support.mailchannels.com/hc/en-us/articles/26814255454093-End-of-Life-Notice-Cloudflare-Workers
I would post this in the community forum, but for some reason access was blocked.
I would post this in the community forum, but for some reason access was blocked.
Help Center
Since launching our free email sending API for Cloudflare Workers customers, we have been proud to provide a simple yet effective solution to deliver messages from Workers code with a minimum of se...
SStarting next week, you may notice a small fraction (1%) of your connections are rejected with an error message containing a link to the same announcement. Any messages that are accepted by the service will be delivered as usual and we will continue to provide support as usual until the final day of service.
JAh, good to hear. We've been advising people to avoid this partnership for quite some time, so it's good to get an official End of Life notice.
JNot sure why you can't access the community, but I've made the announcement there so folks are aware: https://community.cloudflare.com/t/mailchannels-end-of-life-notice-migrate-now/658638
SThanks, @James. I figured you would take up the challenge.
JHappy to help 
JWhat's the best way to cache data from D1 used by a api request via workers?
I need simple data from the D1 database like userCount via a api request, but quering the database on each request doesn't seem optimal.
Is there a way to save the data in some global variable or some caching mechanism so the d1 database isn't queried often?
I need simple data from the D1 database like userCount via a api request, but quering the database on each request doesn't seem optimal.
Is there a way to save the data in some global variable or some caching mechanism so the d1 database isn't queried often?
PHello, is it ok to use the same worker for
scheduledfetch
ZThat is perfectly fine to do. There aren't any real downsides, other than that the two calling methods will both be listed in analytics for the same worker. If your scheduled takes a lot more CPU time, that would throw off your averages for request CPU time usage.
AI noticed new error recently when deploying with
My worker is relatively small with a short list of dependencies (prisma, @planetscale/database, zod), and source map comes out to be around ~7MB. The 5MB limit seems very low. I had to disable source_maps for this worker, but I'd imaging a simple worker that is just proxying calls to a database is going to be a frequent use case of workers.
upload_source_maps=trueMy worker is relatively small with a short list of dependencies (prisma, @planetscale/database, zod), and source map comes out to be around ~7MB. The 5MB limit seems very low. I had to disable source_maps for this worker, but I'd imaging a simple worker that is just proxying calls to a database is going to be a frequent use case of workers.
Kive made a normal worker with eventemitter and it works fine as long as its in the same file, when i put the
bot.on()
Kthe
bot.on()index.jscastam.js
Kas far as i can see eventemitter doesn't work between 2 files...
IIt looks to be because
castam.jsindex.jsindex.jscastamFor it to work the import would need to go the other way around somehow (index.js importing castam.js)
KOhhh i seee
Kokay so now i made a
emitter.js
Kdoesnt work i mean
Kignore the typo in bot.emit() its "hello" there ive fixed it
Ris it best to do a service bindings or the new native rpc calls for worker to worker
WYou'd need to use RPC over service bindings, RPC just provide an interface rather than the worker to worker part
Ryah i re-read the blog from april, makes sense now
ZNeed more RPC and stream r/w examples.FEels like I'm deciphering ancient spells from the few surviving pages of necronomicon
LI'm getting a lot of
Worker exceeded CPU time limitbundledStandardAlso, are there any tips to help figure out what code the CPU limits are caused by?
KLimits only go up
KSubrequests go from 50 to 1,000, CPU time goes from 50ms to 30s, I don't think there's really anything else
KThere isn't really any good solution for it - you can try timing with
Date.now()
LOkay I see, thank you. I recall that there was an issue with
Date.now()asyncI meant for local development, Date.now() isn't limited there
KIn deployed Workers it will still only progress on IO
LAh ok that makes sense. Thanks for the help
KMost Workers only really run into CPU issues when parsing a lot of data or stuff like WASM
LI'm thinking I might be hitting CPU limits with the
HTMLRewrtierI was thinking we could put another worker infront of this worker to detect if there are any issues. Do worker subrequests share the same memory/cpu limits?
KHow can I prevent Cloudflare Workers from counting Unauthorized accesses as requests?
KI am limited to 100,000 requests per day.
KBut it's not fair because someone can easily attack the website and send 100k requests.
KSo hence why I have an API Key so the website can only be authorized/accessed with it.
KBut it seems that Cloudflare Workers counts unauthorized accesses as requests, so someone can easily spam requests to my Workers site.
KAnd hit 100k with spams, and boom, I can't use it anymore.
KIs the API key static or do you look them up in a database somewhere else?
KAPI key is static
Kbut technically i change them manually
KUse the WAF to validate it, and blocked requests won’t reach your Worker