Forge deploy failing

Forge
, , , ,
1

I’m sure it’s something I did - but… when I do forge deploy I get:

ℹ Packaging app files
ℹ Uploading app
ℹ Validating manifest
ℹ Snapshotting functions
ℹ Deploying to environment

Error: Deployment failed

With verbose flag I see

  {
        "description": "Deploying to environment",
        "events": [
          {
            "__typename": "AppDeploymentTransitionEvent",
            "stepName": "Validate hosted resource upload step",
            "createdAt": "2021-09-02T13:18:46.474Z",
            "newStatus": "STARTED"
          },
          {
            "__typename": "AppDeploymentLogEvent",
            "stepName": "Validate hosted resource upload step",
            "createdAt": "2021-09-02T13:18:48.135Z",
            "message": "Deployment failed with unknown error",
            "level": "ERROR"
          },
          {
            "__typename": "AppDeploymentTransitionEvent",
            "stepName": "Validate hosted resource upload step",
            "createdAt": "2021-09-02T13:18:48.137Z",
            "newStatus": "FAILED"
          }
        ]
      }

Help?

2

And of course right after this I send my manifest through a yaml lint checker and the connect fragment is not validating according to it.

Lesson learned - forge lint is not the same between the cli and the deployment mechanism…

3

Well I take that back - not all is well. Managed to get my forge app “installed” but needing upgrading according to the upm - the update button doesn’t work. Forge CLI doesn’t show the installation anymore. Going to the app I get a 500 after I go through the oauth dance…

Works fine if I create a new Confluence instance.

4

I have similar troubles, currently forge install --upgrade is not working:

Error: request to GraphQL Gateway failed, reason: connect ETIMEDOUT 18.184.99.131:443

I’m wondering if some webservicse are down at Atlassian ?

5

Hi @danielwester can I get the forge cli version you were using? You find it by running forge --version.

Can I also request admin access to your site to further investigate? Otherwise, if you could provide me the site url that should be sufficient for now :slight_smile:

1 Like
6

Hey @clement_garin , are you still experiencing the issue?

7

I’m running 2.0.0. I was able to uninstall the forge “proper” app - but the companion connect app still exists.

8

It’s working again on my side, thanks!

9

I think I’ve tracked down my issue to that the name attribute and the remote app key was the same so I’ve ended up with 2 apps installed into an instance with the same key…

10

Hey @danielwester , glad that you were able to track down the issue. Was there anything that lead to apps with the same remote app key being installed?

11

Hey @danielwester ,

2 apps with the same key will do it. Sorry for your bad experience; you’re on the cutting edge here - we’re working on the migration experience right now so that soon, a Forge app will be able to replace a Connect app with the same key, and if you try and install two Forge apps with the same connect key you’ll get a helpful error message.

Let me know if you’d like to set up a meeting to debug connect-on-forge or chat about the internals and our plans for it!

Cheers,
James

12

Hi all, from the reports coming in we suspect there might be an issue with connect-on-forge besides just the inability to rename a key. We’re investigating now. You can follow this ticket for updates.

13

Please see the public issue for an update. It includes a workaround, which is to let us know your app ID so we can let you know your app’s randomly-generated key, in order to set app.connect.key to match.