At present analysis pipelines write all of their logs and errors to text files in the working directory. This is fine, but makes accessing logs complicated, while other messages are dumped in a potentially enormous log file.
We should have a means of accessing logs for an analysis which is independent of scheduler and allows logging events to be accessed through a web interface and the API.
At present analysis pipelines write all of their logs and errors to text files in the working directory. This is fine, but makes accessing logs complicated, while other messages are dumped in a potentially enormous log file.
We should have a means of accessing logs for an analysis which is independent of scheduler and allows logging events to be accessed through a web interface and the API.