`.properties.periods` would throw a `TypeError` if `properties` doesn't exist. For example: ```js co

.properties.periods

.properties.periods

.properties.periods

.properties.periods would throw a

TypeError

TypeError

TypeError

TypeError if

properties

properties

properties

properties doesn't exist. For example:

const obj = {foo: 'bar'}
console.log(obj.foo) // bar
console.log(obj.bar) // undefined
console.log(obj.bar.baz) // TypeError

const obj = {foo: 'bar'}
console.log(obj.foo) // bar
console.log(obj.bar) // undefined
console.log(obj.bar.baz) // TypeError

const obj = {foo: 'bar'}
console.log(obj.foo) // bar
console.log(obj.bar) // undefined
console.log(obj.bar.baz) // TypeError

const obj = {foo: 'bar'}
console.log(obj.foo) // bar
console.log(obj.bar) // undefined
console.log(obj.bar.baz) // TypeError

AA•5/27/23, 7:23 PM

I want to create a single git repo, with lots of workers (mostly webhooks, so stripe/ses/etc) this seems smarter than 1 big project since i can update/change them 1 at a time.. anyone have a good example of how to structure this? just 1 directory per worker, with its own wrangler? or?

JJames `.properties.periods` would throw a `TypeError` if `properties` doesn't exist. F...

IdkWhatever69•5/27/23, 7:24 PM

right right

Screenshot_2023-05-28_at_12.53.54_AM.png

andrew•5/27/23, 7:52 PM

since performance.now()performance.now() isn’t available in Workers and Date.now()Date.now() doesn’t increment unless you do I/O is there any way to actually benchmark CPU bound function execution times in a worker?

kian•5/27/23, 7:58 PM

Nope

kian•5/27/23, 7:58 PM

performance.now()performance.now() is available but it just calls Date.now()Date.now()

andrew•5/27/23, 8:06 PM

That’s annoying. Other platforms mitigate timing attacks but just skewing the resolution, but outright disabling it is an approach i guess

AA•5/27/23, 8:08 PM

wrangler tail <-- all i get are OKs even tho I'm generating internal server errorinternal server error anyplace to look to see better logging?

AA•5/27/23, 8:08 PM

maybe this is a wrangler question..

Walshy•5/27/23, 8:09 PM

what do you mean you're generating it?

Walshy•5/27/23, 8:09 PM

are you catching the exception?

AA•5/27/23, 8:10 PM

not naturally. only if hono is

AA•5/27/23, 8:10 PM

import { Hono } from 'hono'
import { Client } from 'pg'

const app = new Hono()

const connstring = "postgresql://postgres@localhost:5432/hhscdev?schema=public"

app.get('*', async (c) => {
  const client = new Client("postgresql://postgres@localhost:5432/hhscdev?schema=public");
  await client.connect();
  
  const { rows: [{ now }] } = await client.query('select now();');
  return c.text(`Hello Hono! ${now}`)
})

export default app

import { Hono } from 'hono'
import { Client } from 'pg'

const app = new Hono()

const connstring = "postgresql://postgres@localhost:5432/hhscdev?schema=public"

app.get('*', async (c) => {
  const client = new Client("postgresql://postgres@localhost:5432/hhscdev?schema=public");
  await client.connect();
  
  const { rows: [{ now }] } = await client.query('select now();');
  return c.text(`Hello Hono! ${now}`)
})

export default app

Walshy•5/27/23, 8:11 PM

yeah Hono does it's own catching of unknown errors

Walshy•5/27/23, 8:11 PM

So that's why it's ok, the error is being handled

Walshy•5/27/23, 8:11 PM

you can do app.onErrorapp.onError to add your own handler

AA•5/27/23, 8:11 PM

ah, just using that because its recommended. wonder if i should abandon it. or find a way to dump the error

AA•5/27/23, 8:12 PM

really just using a single 'url' per project, so maybe hono is overkill

AA•5/27/23, 8:20 PM

@walshydev thanks. 1 problem down. 99 more to go

AA•5/27/23, 8:23 PM

ooh, another minor quirk. i'm doing a wrangler --env staging.. but in the workers dashboard it says the environment is 'production' are these things not linked?

AA•5/27/23, 8:23 PM

also did a vars = { ENVIRONMENT = "staging" }

Cyb3r-Jak3•5/27/23, 8:24 PM

You probably have two workers with different names and both using the production environment

AAA ooh, another minor quirk. i'm doing a wrangler --env staging.. but in the worke...

JamesOP•5/27/23, 8:24 PM

The terminology around environments is weird. I summarised a while back: https://community.cloudflare.com/t/environments-vs-deployments/451036/2?u=cherryjimbo

AA•5/27/23, 8:25 PM

@cherryjimbo Oh you're CherryJimbo! that explains so much

also this explains lots too. just ignore the dashboard one

JamesOP•5/27/23, 8:26 PM

Yep that's me

yeah, you're best just to ignore the productionproduction stuff on the dashboard entirely right now

JamesOP•5/27/23, 8:26 PM

This'll get better with Pages+Workers convergence long-term, but that's probably a ways off

AA•5/27/23, 8:26 PM

https://developers.cloudflare.com/workers/platform/environments/#introducing-environments <-- i'm setting up my workers like this, is this reasonable?

Environments · Cloudflare Workers docs

Environments are different contexts that your code runs in. The Workers platform allows you to create and manage different environments. Through …

AA•5/27/23, 8:26 PM

change the 'name' per environment

JamesOP•5/27/23, 8:27 PM

yep, that's basically the only way to do it right now

AA•5/27/23, 8:27 PM

i see ok

AA•5/27/23, 8:28 PM

now to figure out postgres through cloudflareD without exposing it to the internets

AA•5/27/23, 8:36 PM

maybe not possible..

ItsWendell•5/27/23, 9:45 PM

Does anyone here has a comfortable setup with Toucan.js (Sentry) and SourceMaps?

Walshy•5/27/23, 9:49 PM

yep but not one I can easily share, the repo has a good example setup though

Walshy•5/27/23, 9:49 PM

https://github.com/robertcepa/toucan-js/tree/master/examples/wrangler-basic

kian•5/28/23, 1:23 AM

The only thing that I know of which has accurate duration on a per-request basis is the HTTP Requests dataset on Logpush - which is an Enterprise feature

kian•5/28/23, 1:23 AM

Workers for Platforms will probably have better metrics, as customer Workers is one of the main use-cases for it, but that's also currently Enterprise but is likely to come to pay-as-you-go plans eventually.

kian•5/28/23, 1:23 AM

Timing in Workers is always difficult as every timer you have accurate to (including the timestamps in the live logs) implements Spectre mitigations.

kian•5/28/23, 1:24 AM

It only advances on I/O - so it's never representative of wall-clock anyways.

SoulHarsh007•5/28/23, 2:08 AM

is there a way I can get execution region from my worker?

Walshy•5/28/23, 2:09 AM

request.cf.colorequest.cf.colo

Walshy•5/28/23, 2:10 AM

or request.cf.regionrequest.cf.region for more region like info

Walshy•5/28/23, 2:10 AM

for details on all the available properties: https://developers.cloudflare.com/workers/runtime-apis/request/#incomingrequestcfproperties

Walshy•5/28/23, 2:12 AM

there are no "regions" in CF though so yeah meowupsidedown

SoulHarsh007•5/28/23, 2:20 AM

And this works in a cron job too?

SoulHarsh007•5/28/23, 2:22 AM

I mostly want it for my cron jobs, since I want to know which region executed the cron. I was also wondering if it might be possible to run the same cron job, from multiple regions at the same time?

Walshy•5/28/23, 2:28 AM

err no in cron we don't expose that easily but you can hit /cdn-cgi/trace to grab it

const res = await fetch('https://cloudflare.com/cdn-cgi/trace');
const text = await res.text();
const colo = text.substring(text.indexOf('colo=') + 5, text.indexOf('\n', text.indexOf('colo=')));

const res = await fetch('https://cloudflare.com/cdn-cgi/trace');
const text = await res.text();
const colo = text.substring(text.indexOf('colo=') + 5, text.indexOf('\n', text.indexOf('colo=')));

`.properties.periods` would throw a `TypeError` if `properties` doesn't exist. For example: ```js co

Similar Threads

Similar Threads

Similar Threads