Vercel creates failed pipelines in Gitlab

[kspworks (Kasper)]

I have a NextJS repository in GitLab, which is deployed to Vercel by linking GitLab account to Vercel and selecting that repository.

Initially, all my branches (main and dev), and all merge requests were deploying to Vercel (I assume it is default behaviour), but over time I decided that I only want deployments of main branch to be deployed to Vercel, so I disabled “preview deployments” in Vercel project settings, and disabled dev branch deployments in vercel.json file.

And while MRs and dev branch stopped deploying to Vercel (which is what I wanted), I got an issue - when I push new changes to GitLab’s dev branch, for some reason after about 25 minutes a new pipeline appears (in GitLab) for dev branch, instantly in “failed” state, and GitLab sends me an email about failed pipeline, which is annoying.

I am the single maintainer in this repo, no one else has access to the code and/or my Vercel project, so manually-triggered Vercel deployments are off the table (Vercel chatbot asked about it).

No dev branch deployments are appearing in Vercel Deployments page of this project.

I also have a .gitlab-ci.yml file (for deployments to different hosting provider), but it does not contain any deployment config regarding Vercel (the issue with failed dev pipelines occured long before this file was created).

I tried to reconfigure Vercel webhook in GitLab settings to only trigger on main branch, but this didn’t fix the issue, failed pipelines are still appearing every time I push to dev branch.

My vercel.json file contents (it is placed in the root directory of my repo):

vercel.json

{
  "crons": [
    {
      "path": "/api/cron/daily-0000",
      "schedule": "0 0 * * *"
    },
    {
      "path": "/api/cron/daily-0900",
      "schedule": "0 9 * * *"
    }
  ],
  "git": {
    "deploymentEnabled": {
      "dev": false
    }
  }
}

My .gitlab-ci.yml file contents (also placed in the root):

.gitlab-ci.yml

image: alpine

before_script:
  - 'which ssh-agent || ( apk add --update openssh-client git )'
  - eval $(ssh-agent -s)
  - ssh-add <(echo "$SSH_PRIVATE_KEY_BASE64" | base64 -d)
  - mkdir -p ~/.ssh
  - chmod 700 ~/.ssh
  - '[[ -f /.dockerenv ]] && echo -e "Host *\n\tStrictHostKeyChecking no\n\n" > ~/.ssh/config'

workflow:
  rules:
    - if: $CI_COMMIT_BRANCH == "main"
      variables:
        BUILD_PATH: $PROD_BUILD_PATH
        OUT_PATH: $PROD_OUT_PATH
      when: always
    - when: never

stages:
  - deploy

production:
  stage:
    deploy
  environment:
    name: production
  rules:
    - if: ($CI_COMMIT_BRANCH == "main") && $BUILD_PATH && $OUT_PATH && $SSH_CONNECT
      when: on_success
    - when: never
  script:
    - ssh $SSH_CONNECT "[REDACTED]"

Vercel Git settings:

Vercel.png

This is how failed pipelines appear in GitLab UI:

Pipelines.png

And this is how Vercel webhook configured in GitLab:

Webhook.png

I know there’s an option to disable Vercel’s deployments entirely, and move off to GitLab pipelines which would deploy to Vercel manually, but for me this is a last resort option because I’d like to keep more GitLab CI/CD minutes for my other projects.

So, does anybody knows a way to get rid of those “failed” pipelines of dev branch in GitLab?

[jacobparis]

Hi Kasper, I’ve escalated this internally and the team is aware of it. They said they should have a fix out in the next few days, I’ll follow up with you again once that ships

Thanks for your patience, the extra notifications should go away soon

[jacobparis]

The underlying bug here has been fixed internally so this should be good to go now.

Give it a shot and let me know if you’re still getting failed pipelines so I can put the team back on it

[kspworks (Kasper)]

Can confirm that the issue is fixed, failed pipelines no longer appear in GitLab after pushing to dev branch.

Thanks a lot to you and Vercel team, and have a nice day!