• Icon: Improvement Improvement
    • Resolution: Fixed
    • Icon: Minor Minor
    • core
    • None
    • 2.434

      I moved our Jenkins to a new server a few weeks ago and since it doesn't actually do anything, I gave the VM only few resources (the agents have the workload). This led to the main node being marked as offline, where it looked like this:

       

      Finding the cause was a bit time consuming, which would actually be easy to fix:

      • Within the table Nodes the icon has no tooltip (not like the icon within the build executor list)
      • Disk space is too low. Only <0>GB left on <1>. (DiskSpaceMonitorDescriptor.DiskSpace.FreeSpaceTooLow) isn't this much helpful, as you don't know, how man disk space is required. So the test should also include the configured value of Free Space Threshold.
      • As we running Jenkins already for years I didn't knew where I can change the settings, so I had to use a search machine. Node Monitoring sound more like sth. like Grafana or like Load statistics of a node. So maybe renaming it to Node Monitoring Settings would be useful
      • Marking the node as Offline is rather confusing, as to me that sounds more like losing connection to the nodes. Defective would be better, but maybe someone has better ideas.
      • An earlier warning would be useful. For example, if you fall below 20 GB, there is a warning that the disk is running low. And only at 10 GB the node is deactivated.

          [JENKINS-72009] Improve warning when disk space is to low

          Just a quick note: it's no longer recommended to run anything on the built-in node.
          https://www.jenkins.io/doc/book/managing/nodes/#components-of-distributed-builds 

          Bruno Verachten added a comment - Just a quick note: it's no longer recommended to run anything on the built-in node. https://www.jenkins.io/doc/book/managing/nodes/#components-of-distributed-builds  

          Markus Winter added a comment -

          The problem I see is that it is not possible to define the limits individually per agent currently (I'm not aware of a plugin that can do this).  When the complete disk of the controller is < 50GiB it probably doesn't make sense to have a threshold of 20GiB, but then your agents run huge builds so you need a disk of 1TiB, there it might make sense to have a threshold of 20GiB or even more.

          Markus Winter added a comment - The problem I see is that it is not possible to define the limits individually per agent currently (I'm not aware of a plugin that can do this).  When the complete disk of the controller is < 50GiB it probably doesn't make sense to have a threshold of 20GiB, but then your agents run huge builds so you need a disk of 1TiB, there it might make sense to have a threshold of 20GiB or even more.

          Markus Winter added a comment -

          Partially duplicated by JENKINS-24947

          Markus Winter added a comment - Partially duplicated by  JENKINS-24947

            mawinter69 Markus Winter
            sunblack sunblack
            Votes:
            0 Vote for this issue
            Watchers:
            3 Start watching this issue

              Created:
              Updated:
              Resolved: