You are already using it in production? - Cloudflare Developers

© 2026 Hedgehog Software, LLC

More

Communities Docs About Terms Privacy

C

cjOP•4/9/24, 2:54 PM

Are there any limitation on the number of tabs opened? Memory limit or something?

I know there is none stated in the limits page but did you face any issue while in production?

Ccj Are there any limitation on the number of tabs opened? Memory limit or something...

rob

R

rob•4/9/24, 5:02 PM

Im curious on that also

Ccj You are already using it in production?

Advany

A

Advany•4/9/24, 7:43 PM

I had a hard time crashing it. I set the limit to 25 for myself to be save.

Also using it in production.

I did test it once with 100 to 1000 current pages. Crashed around 200ish.

Advany

A

Advany•4/9/24, 7:44 PM

I did page to pdf and page to screenshot and rxtracting page dom

rob

R

rob•4/9/24, 9:01 PM

on all 200 ?

jason

J

jason•4/10/24, 1:19 PM

I don't know whose job this is, but there should be a space in the title

CleanShot_2024-04-10_at_08.18.242x.png

Jadugar_Jaggu

J

Jadugar_Jaggu•4/10/24, 6:59 PM

Kindly add dark mode on this domain also https://blog.cloudflare.com/ , my eyes are burning

Chaika

C

Chaika•4/10/24, 8:53 PM

this is the channel for browser rendering api. The community forums have a generic feedback category

Advany

A

Advany•4/10/24, 9:56 PM

Whats the timeline for expanding the limits for the amount of sessions allowed?

Advany

A

Advany•4/10/24, 9:57 PM

Anyway to request higher limits? Happy to pay!

Chaika

CChaika this is the channel for browser rendering api. The community forums have a gener...

Jadugar_Jaggu

J

Jadugar_Jaggu•4/11/24, 3:53 AM

If I execute a Javascript function on a webpage using browser rendering and it will take too long to complete it so in this situation what will be the behaviour of a Cloudflare worker, like it will wait for the completion of that function which is been executed in browser session or it will simply kill itself due to some kind of timeout error or we can leave the browser session open and doing the work and leave the worker?

scythe

Sscythe If the browser does not receive a command every minute at least, it will shutdow...

Jadugar_Jaggu

J

Jadugar_Jaggu•4/11/24, 5:42 PM

interesting , it mean we can force a browser session to stay open as long as we want

dave

D

dave•4/11/24, 8:49 PM

Did this become an open beta?

James

J

James•4/12/24, 1:35 AM

It’s now generally available to all workers paid plans

Vikash

V

Vikash•4/12/24, 7:47 AM

Can I connect the same browser in multiple workers? I see the getRandomSessiongetRandomSession function in the reuse browser example, which filter out the existing connected browser sessions using connectionIdconnectionId

const sessionsIds = sessions
            .filter(v => {
                return !v.connectionId; // remove sessions with workers connected to them
            })
            .map(v => {
                return v.sessionId;
            });

const sessionsIds = sessions
            .filter(v => {
                return !v.connectionId; // remove sessions with workers connected to them
            })
            .map(v => {
                return v.sessionId;
            });

But, in a worker I am just capturing a screenshot using queue()queue() with single page and closes that immediately. So I should be able to re-use the browser in multiple worker as there might be 10-15 pages without any issue I guess in same browser. Or it's limited to single connection only and I will get error if I try to use a browser which is currently in use on any other worker?

Ssc Yes but no eta yet.

Vikash

V

Vikash•4/15/24, 10:54 AM

I tried to use the reuse session example to capture screenshot for 100s of url, the page.gotopage.goto randomly freeze and worker stops processing any more URL from queue()queue() or 15 minutes

Vikash

V

Vikash•4/15/24, 10:55 AM

Vikash

V

Vikash•4/15/24, 10:57 AM

even when I have timeout set explicitly in page.goto

const response = await page.goto(url, {
timeout: 15000,
waitUntil: "domcontentloaded",
});

Ssc Check that you're cleaning up resources, eg. closing pages etc. Random freezes s...

Vikash

V

Vikash•4/15/24, 11:07 AM

The puppeteer has 30000 ms by default on launch, so is should timeout. instead the worker goes unresponsive for 15 minutes

Ssc Yes but no eta yet.

Advany

A

Advany•4/15/24, 2:11 PM

Any eta for:

Customizing session timeout (i saw a branche for that on github)
Getting more sessions?

Ssc Check that you're cleaning up resources, eg. closing pages etc. Random freezes s...

Advany

A

Advany•4/15/24, 2:34 PM

whats the method to check for memory usage?

Jadugar_Jaggu

J

Jadugar_Jaggu•4/16/24, 8:52 AM

I think browser.pages() is not able to detect how many pages are open in the current session. Is it a known bug?

Jadugar_Jaggu

J

Jadugar_Jaggu•4/16/24, 8:53 AM

I have one page which is running currently but when I am executing the browser.pages() on same session its returning me a empty array ( length 0 )

Ssc For 1. Should be in less than a month. For 2. No eta yet

Advany

A

Advany•4/16/24, 11:14 AM

awesome! 1 will help a lot!

Jadugar_Jaggu

J

Jadugar_Jaggu•4/16/24, 3:58 PM

we implemented a cron trigger in the worker which is using browser-rendering api to execute specific commands to ensure session continuity. However, as shown in the screenshot, the connection that started at 4:42:23 will end at 4:42:25. Just three seconds after the end, the browser closes at 4:42:28, citing idle as the reason.

According to the documentation, A browser instance gets killed if it does not get any command for 60 seconds, freeing one instance.A browser instance gets killed if it does not get any command for 60 seconds, freeing one instance.

Brian

B

Brian•4/16/24, 4:36 PM

Im testing the browserless api with the example screenshot worker. I cant get it to work when accessing a map tile url. It seems webgl is not supported by chromium. Has anyone been able to find a workaround for this? My hope is to use a browserless worker to make screenshots of maps for users.

Brian

BBrian Im testing the browserless api with the example screenshot worker. I cant get i...

Jadugar_Jaggu

J

Jadugar_Jaggu•4/16/24, 4:55 PM

may be you can try using Google maps lite mode
https://www.google.com/maps/?force=litehttps://www.google.com/maps/?force=lite

http://maps.google.com/?force=canvashttp://maps.google.com/?force=canvas

Brian

B

Brian•4/16/24, 5:06 PM

Thanks but I'm trying to grab a screenshot of my own tileservice that use maplibre to render vector tiles to the browser. Maplibre uses webgl and when I grab a screenshot of the url I get a white blank screenshot.
I saw others ask about using webgl in browserless workers and it seems it hasnt been solved yet.

Mercury

M

Mercury•4/18/24, 9:27 AM

It gives me error text: 'workers.api.error.no_access_to_browser_worker [code: 10088]'text: 'workers.api.error.no_access_to_browser_worker [code: 10088]' Is this still in Private Beta?

Mercury

M

Mercury•4/18/24, 11:15 AM

Hi team, I am getting this on screenshot. Any idea how we can resolve this?

Screenshot_2024-04-18_at_16.45.26.png

Mercury

M

Mercury•4/18/24, 11:57 AM

Hi team, I am also not able to render canvas in Browser Rendering.

villainy

V

villainy•4/18/24, 8:24 PM

getting this error:

 ⨯ Error: Failed to load external module node:stream: TypeError: Native module not found: node:stream

 ⨯ Error: Failed to load external module node:stream: TypeError: Native module not found: node:stream

villainy

V

villainy•4/18/24, 8:24 PM

trying to use with nextjs

villainy

V

villainy•4/18/24, 8:25 PM

compatibility_flags = [ "nodejs_compat" ]
node_compat = true

compatibility_flags = [ "nodejs_compat" ]
node_compat = true

villainy

Vvillainy ``` compatibility_flags = [ "nodejs_compat" ] node_compat = true ```

Chaika

C

Chaika•4/18/24, 8:31 PM

you can't use both, one or the other

villainy

V

villainy•4/18/24, 8:32 PM

getting the same error regardless

villainy

V

villainy•4/18/24, 8:32 PM

this didnt help either

import { Stream } from "node:stream";
globalThis.Stream = Stream;

import { Stream } from "node:stream";
globalThis.Stream = Stream;

Chaika

C

Chaika•4/18/24, 8:33 PM

node:stream is only a thing in the compat flag version. I don't think that's the typical error though, "Native module not found"?
I would make sure you're on the latest cloudflare/puppeteer at least, the first versions used node_compat rather then the flag

villainy

V

villainy•4/18/24, 8:34 PM

on the latest version

    "@cloudflare/next-on-pages": "^1.11.0",
    "@cloudflare/puppeteer": "^0.0.6",

    "@cloudflare/next-on-pages": "^1.11.0",
    "@cloudflare/puppeteer": "^0.0.6",

villainy

V

villainy•4/18/24, 8:35 PM

node:

v20.10.0

villainy

V

villainy•4/18/24, 8:36 PM

maybe i need to downgrade to 18?

const sessionsIds = sessions
            .filter(v => {
                return !v.connectionId; // remove sessions with workers connected to them
            })
            .map(v => {
                return v.sessionId;
            });

const sessionsIds = sessions
            .filter(v => {
                return !v.connectionId; // remove sessions with workers connected to them
            })
            .map(v => {
                return v.sessionId;
            });