Hello,

Given you now have the reference variables set correctly now, I'm going to take a guess and say that you deployed browserless into its own separate isolated project?

Yes! Sorry, should have updated. Created the template and copied stuff over. Thanks so much!

Awsome!

Need help with anything else?

Yes, actually! Browserless actually doesn't seem to be working as expected.

Basically, the arch of my app is that I have a server which queues a bunch of scraping / Browserless tasks using RabbitMQ, and then a bunch of consumers (I'll use Railway replicas when I scale this up) which actually complete them.

My consumers work perfectly locally, but in Railway no luck. I've tried manually copy-pasting the resolved ENV variables from the Browserless template, and that doesn't seem to do it.

For example, working version locally:

Non-working version on railway — you'll note, consumes tasks perfectly, just seems like a Browserless problem.

PS — Railway is really amazing! And I've been using a ton of your templates, thanks so much for all of this! Makes it way easier to hack on side projects without having to manage a bunch of infra

Are you using Browserless locally?

Yes I am! I'm running the exact stuff I'm using on Railway locally / without problem

Are you running the same Browserless version on Railway? there are two templates for Railway, Browserless v1 and v2

Hmm...I'm using browserless/chrome:latest locally I believe, although I seem to be connecting to browserless in a version-agnostic way. I think I might be a bit confused about how the template work here. My assumption was that I need the Browserless template running so that I can pass the BROWSERLESS_WEBDRIVER_ENDPOINT to my regular / vanilla node server, and it can make calls to browserless. (I.e. this is the only part of teh template I'm using). I was using the Browserless v2 template, but tried with V1 (copy-pasting the Raw Editor env variables) and also no luck. Any ideas?

(Thanks so much for help here, just not quite sure what is going on!)

On Railway you are using ghcr.io/browserless/chromium so thats browserless v2 that uses Chromium.

There are some examples repos here -

https://railway.com/template/0jqemX

If your library and language isn't there, I can try to add it.

Ok wait, i think i must be confused!

Library / language almost certainly is there - I'm just using JS! Browserless v2 should be running, and I'm using all of your default env variables, I believe. For some reason, my node backend can't seem to use browserless - why would this be? I am almost certain this is a railway networking / setup thing, although also possibly unrelated user error (but I think unlikely, because all working locally! The only issue is getting local to run on Railway infra

Have you successfully used Browserless hosted on Railway with locally running code?

I did earlier, but I'm not seeming to have some trouble! A few things I did:
- Copied your Selenium / Node example, put it into railway with a few of my changes (I had it start to screenshot my website / upload it to Cloudinary as a test). Attached error, not sure exactly what's going on, but it's basically the boilerplate.
- Added some logs to my local code (that I believe I started with the template). To confirm: I have my Node project that is NOT the Browserless template (it's a RabbitMQ consumer), and separately, the Browserless template deployed. Theoretically the Node consumer can call / control the Browserless instance, but in practice no.

(Also, separately, seems like Browserless does queueing, and I maybe didn't need this. Either way though, I can't get just vanilla Browserless running as is, so maybe something else is wrong / something with my configuration is wrong?)

Thank you for your help!

You are using Browserless v2, v2 has dropped support for Selenium, you need to be using v1.

Ah! Thank you, didn't realize that! Perfect

Working now?

Nope, no luck sadly. I:
- Swapped Browserless for v1
- Swapped the BROWSER_TOKEN and BROWSER_WEBDRIVER_ENDPOINT on each of the services that needs to access browserless to the stuff generated by Browserless (is this correct? btw this manual process, if so, is fairly clunky - i.e. have to manually reveal the endpoints on the Browserless service and copy over, and can't use the raw editor because they are template variables)

Getting this now:

You can just click the + New Variable button, start typing and it will autocomplete for you, no need to manually copy anything.

Are you able to run this code locally by having it connect to the Browserless instance hosted on Railway?

You have also manually modified the variables the Browserless template comes with, I'm not sure why that was done?

Ok I think I see!

I actually just started an unrelated project for work that has similar functionality (the former was a personal proj)

I guess I'm confused:
- do I deploy a Browserless template and then separately deploy my service?

OR

- do i deploy a Browserless template, and then change the source Git repo to my version of the selenium example?

I tried the first way, and copy-pasted all of the ENV variables from the raw editor to my service. When I hit the endpoint externally, I get:
```
{

"error": "Invalid URL: https:///webdriver"

}
```

When I try the second way, it doesn't re-deploy on pushes, seems like the first is correct?

Sorry for all of this back and forth, hoping to figure out both asap!

(This one, confirmed, works locally)

(Also separately just want to confirm, there's "iron wall" separation between my personal projects (on my account) and my team projects. Want to make sure company doesn't get billed for my personal usage (or vice versa)

You would want to deploy Browserless and then your own app in a separate service.

Please do not copy and paste any variables, that url is incorrect, use the reference variables by letting it autocomplete.

Personal and Team workspaces have completely separate billing, yes.