Unti now, we’ve been deploying our service following successful tests using the ‘deploy’ directive in circle.yml. This is fine until it breaks. Generally it’s bug-free, it’s likely to break due to external factors, which we work to resolve and then rebuild. Of course rebuilding means waiting for the entire test run to occur again before getting to the deployment state where it could potentially fail.
We’re breaking out the build in to a workflow to simplify the process representation in the Circle UI, however one feature I’d have hoped existed as part of the workflows is to rebuild from a specific job onwards. So in the previous case, we’d like to be able to resolve the external deployment issue, leaving in tact the passed unit tests in the workflow and go straight to re-running the previously-failed deployment job (and then automatically running subsequent jobs if required). Presumably the persistence data would remain in tact still too.
If this isn’t a feature yet, it would make life much easier for us! Ultimately, if we could somehow make build modifications to a workflow and it re-run the changed jobs only, that would be something pretty special.