Add retry parameter for smoke tests.

### 📝 Description of the feature

Since lately we have been getting more and more transient connectivity errors, do you think it is a good idea to add a retry parameter to the `build_wheelhouse` action? This way we might avoid to manually rerun the action if it was failing due to some transient connectivity issue.

### 💡 Steps for implementing the feature

_No response_

### 🔗 Useful links and references

_No response_

Locate the build_wheelhouse action and read how its inputs and failure handling are defined. Add the requested retry parameter and verify that transient connectivity failures are retried while successful runs remain unchanged.

Written by the indexing model from the issue text.