Jobs failure and dashboard
Related to #280 (closed)
Currently only jobs that encounter a run time error are marked as failed on the dashboard
However, some fetchers have the very desirable feature to raise and capture an exception in case of download failure or parsing error, to log the error and to continue with the next dataset. This behavior isn't recognized by the dashboard. See for an example #276 (closed)
We need to devise a mechanism where the fetcher returns a non OK status in case of logged error so as to be captured on the dashboard