dmorfav opened a new issue, #9370:
URL: https://github.com/apache/devlake/issues/9370

   ### Search before asking
   
   - [X] I had searched in the 
[issues](https://github.com/apache/devlake/issues?q=is%3Aissue) and found no 
similar issues.
   
   ### What happened
   
   When adding data scopes to an **Azure DevOps Go** connection, repositories 
hosted outside Azure Repos (Bitbucket, GitHub) that are built by Azure 
Pipelines do not show up at all. Only `TfsGit` repositories are listed.
   
   In `backend/plugins/azuredevops_go/api/remote_helper.go`, `listRemoteRepos` 
calls `GetRemoteRepositories` once per service endpoint of a supported type and 
waits for all of them with an `errgroup`. If **any single** endpoint call fails 
or times out, the whole result is discarded (`"failed to call 
'GetRemoteRepositories', falling back to empty list"`), so none of the external 
repositories are listed.
   
   In an Azure DevOps project with many Bitbucket/GitHub service connections 
(in our case around 60, some of them stale or slow), at least one call almost 
always fails. External repositories therefore never appear in the UI, even 
though the endpoint that actually backs the pipeline works fine.
   
   ### What do you expect to happen
   
   A failing service endpoint should be skipped (and logged), and the 
repositories from the endpoints that respond should still be listed.
   
   ### How to reproduce
   
   1. In an Azure DevOps project, have a pipeline whose source repository is on 
Bitbucket (or GitHub), plus several Bitbucket/GitHub service connections, at 
least one of which fails or times out when listing repositories (e.g. expired 
credentials).
   2. In DevLake, create an Azure DevOps Go connection and browse that project 
to add data scopes.
   3. Only Azure Repos (`TfsGit`) are listed; the external repositories are 
missing.
   
   ### Anything else
   
   Workaround: the scope can be created directly with `PUT 
/plugins/azuredevops_go/connections/{connectionId}/scopes`, using `type: 
"Bitbucket"` and `ExternalId: "<workspace>/<repo>"`. After that, builds are 
collected and production deployments are detected as expected, which suggests 
only the listing step is affected.
   
   Possible fix: in `listRemoteRepos`, log per-endpoint errors and keep 
collecting results instead of returning an error from the errgroup, and/or 
apply a per-endpoint timeout.
   
   ### Version
   
   v1.0.3-beta19 (the same code is on `main`)
   
   ### Are you willing to submit PR?
   
   - [ ] Yes I am willing to submit a PR!
   
   ### Code of Conduct
   
   - [X] I agree to follow this project's [Code of 
Conduct](https://www.apache.org/foundation/policies/conduct)
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to