Optimize jenkins artifact extraction

The jenkins analysis task is being fixed in #399, but the fix is kind of ugly because it requires traversing the entire filesystem in order to extract the right files. We should find a better solution and/or fix the upstream file globbing in Plaso (either the filter file syntax or the artifacts syntax).

Start by reading the Jenkins analysis task and the fix referenced in issue #399. Investigate how Plaso file globbing handles its filter-file and artifacts syntax, then compare that with the current full-filesystem traversal. Done means artifact extraction avoids the broad traversal while still selecting the required files.

Written by the indexing model from the issue text.