Hey all, relatively new developer working on a forge application that will be using external Oauth connecting to my companies IdP and using external api within jira.
ERROR 22:00:48.753 ba72d2d9e507131d Fetch failed for remote ‘myIpro-IdP’, provider ‘myIpro’, path ‘/connect/userinfo’ no credentials previously requested
is the error I have received.
The above is the code in the main app index.js. I believe the actions and remotes have all been set up correctly in the manifest, I have provided the correct permission scopes namely read:user:jira, as well as
defined external fetch client and backend scopes to include the url to the remote.
what am I missing? I have been following along with the guide on Oauth for google apis and modifying this to suit my companies IdP provider.
It seems that other users have found that this error only happens due to the tunnel being active?
Im not so sure. Following along with the google example I used the retrieveProfile action in my fetch. here are my other actions:
That “error” is normal, and is what happens when you call requestCredentials(). It raises an exception that is caught by the runtime to show the button.
Other than the error log line, what else is happening?
After that requestCredentials method is run it looks like it just fails. I have console statements after that call that dont run, and when I check our companies IdP logs we dont log any sort input / request from our Jira client, so it doesnt seem to be sending the request to auth like it should be.
Part of the issue indeed was my function for the resolver was missing the provider, adding this yesterday allowed my companies IdP to begin to log. It looks like its establishing auth correctly on our side
Hi @JasonDunn, thanks for persisting with this issue you are having!
That last screenshot shows an error is happening during the exchange step (where the system exchanges the authorization token for an access token).
From the manifest screenshot you provided before, this would be an issue talking to the /connect/token endpoint.
From the logs on our side, I can see it is getting a HTTP 503 response back, suggesting an issue with a load balancer, or the service being down.