Add Service Template: Velld

[ejscheepers]

Changes

• Added Service Template: Velld

[github-actions]

Hi @ejscheepers! 👋

It appears to us that you are either adding a new service or making changes to an existing one.
We kindly ask you to also review and update the Coolify Documentation to include this new service or it's new configuration needs.
This will help ensure that our documentation remains accurate and up-to-date for all users.

Coolify Docs Repository: https://github.com/coollabsio/coolify-docs
How to Contribute a new Service to the Docs: https://coolify.io/docs/get-started/contribute/service#adding-a-new-service-template-to-the-coolify-documentation

[ejscheepers]

Thanks @ShadowArcanist - I made the requested changes, will do a PR to add healtcheck to Velld Web App.

Will push commit as soon as Github outage is fixed..

[ShadowArcanist]

Hey @ejscheepers thanks for the PR and following along with implementing requested changes!

Unfortunately I have to hold this PR because the app is unstable (check attached video):

Velld.CY.mov

• Clicking "discover" option will throw a client side exception then restarting the app will break the entire app by showing bad gateway. I restarted it many times but it didn't fix the issue and the app logs is not showing any error messages.
• This behavior is consistent, I redeployed the app over 7 times and every single time it breaks like shown in the video above.
• If the team behind this releases a fix then please test this app on the fixed version, once you have confirmed everything is working then update this PR and request a review from me.

[ejscheepers]

@ShadowArcanist I will submit a PR to the project to fix the error that was crashing on DB discovery.

As for the restart resulting in "Bad Gateway" I will have to investigate a bit more, because on my NON-COOLIFY docker compose it works perfectly, so maybe Coolify is doing something weird to the compose that's breaking it?

[ShadowArcanist]

@ShadowArcanist I will submit a PR to the project to fix the error that was crashing on DB discovery.

As for the restart resulting in "Bad Gateway" I will have to investigate a bit more, because on my NON-COOLIFY docker compose it works perfectly, so maybe Coolify is doing something weird to the compose that's breaking it?

Thanks!

I don't think Coolify is doing anything (I could be wrong here) to cause that Bad Gateway the container don't have a healthcheck so Coolify proxy should route requests properly and the container logs doesn't throw any error.

The compose files used by Coolify can be found on /data/coolify/services/<uuid_of_service>/docker-compose.yml (might be helpful for you to debug)

[ejscheepers]

@ShadowArcanist I think I've found the issue:

Root cause
Coolify injects the public URLs via SERVICE_URL_API_8080 and SERVICE_URL_VELLD_3000.
On redeploy it generates a new resource UUID (new subdomain), but keeps the old values from the previous deployment in the generated compose.
The Go backend starts with a stale public URL → generates invalid absolute links → many requests fail → Traefik has no healthy upstream → 502.
Fix
Explicitly override the variables with an empty value in the source compose:

services:
  api:
    environment:
      SERVICE_URL_API_8080=    # empty = forces Coolify to inject the correct fresh URL every time
  web:
    environment:
      SERVICE_URL_VELLD_3000=  # same for the Next.js frontend

Not sure if this is a feature or a bug, but this is what was breaking it..

[ShadowArcanist]

@ShadowArcanist I think I've found the issue:

Root cause Coolify injects the public URLs via SERVICE_URL_API_8080 and SERVICE_URL_VELLD_3000. On redeploy it generates a new resource UUID (new subdomain), but keeps the old values from the previous deployment in the generated compose. The Go backend starts with a stale public URL → generates invalid absolute links → many requests fail → Traefik has no healthy upstream → 502. Fix Explicitly override the variables with an empty value in the source compose:

services:
  api:
    environment:
      SERVICE_URL_API_8080=    # empty = forces Coolify to inject the correct fresh URL every time
  web:
    environment:
      SERVICE_URL_VELLD_3000=  # same for the Next.js frontend

Not sure if this is a feature or a bug, but this is what was breaking it..

We had a bug which makes Coolify not update value of env that contains port at the end when user updates the URL - this was fixed on v4.0.0-beta.445.

I just tested the app on v4.0.0-beta.446 and the Envs have correct values but still it shows the bad gateway error:

Also Coolify does not automatically change the auto generated url or the resource uuid - it's set on creation of the service, and Docker compose empty or one click services doesn't have redeploy option so you either restart or pull latest image and restart

So are you sure it's caused by Coolify? asking because it's still not working for me even when the ENV are correct for both api and nextjs frontend

Note: Coolify v4.0.0-beta.445 and v4.0.0-beta.446 were not released to public yet.. I am testing the service on dev instance

[ejscheepers]

This is very weird, I tested with empty docker compose on Coolify [v4.0.0-beta.444] and it worked on my side, even after restarting.