Hi Frederic and Mikael,
    I saw several build jobs failed with
    [ERROR] Plugin some.group.id:some-artefact:pom:1.2.3 or one of its dependencies could not be resolved:
        Failed to read artifact descriptor for some.group.id:some-artefact:pom:1.2.3:
        Could not transfer artifact some.group.id:some-artefact:pom:1.2.3 from/to central
        (https://repo.maven.apache.org/maven2): Connection reset -> [Help 1]
    With current setup, each job always downloads all the artifacts
      from maven central again and again. This looks like central
      servers are refusing requests from eclipse.org from time to time.
     Did you consider some way of caching?
    1. there can be some local nexus cache on your machines which can
      speed up downloading and reduce number of central download
      requests
    2. some permanent workspace for each job with maven repository
      (~/.m2/repository) pointing there may reduce job download requests
      as well
    
    
    Today my https://jenkins.eclipse.org/metro/job/wsit-ee4j8-build/
      failed for 3rd time in a row with this error, but each time it was
      different artifact in random phase of the build. That's why I
      think we need at least some way of artifacts caching.
    
    
    Tomas