Agenda

1. Item 1

Minutes

• Attendees:
  • VO - Person 1

1. Tutorial on the Lemon monitoring system by Ivan Fedorko.
2. Presentation on current Lemon metrics and exceptions definitions used by Atlas to monitor file system usage, certain service daemons being in the process table(, etc) by Flavia
   1. Slides of these two presentations are available in Indico:
      http://indico.cern.ch/categoryDisplay.py?categId=3006
3. Discussion
   1. Experiments other than Atlas (LHCb, CMS, ..) are also configuring lemon metrics and exceptions (and even developing their own lemon sensors) for monitoring the services running on their VOBoxes.
   2. There is no central index which contains all lemon sensors, but central ones are all in a central repository (in SVN). Developers of new lemon sensors can request their new sensors to be added to the central repository by contacting the lemon support team.
      1. It was felt that a central index for VO-shared lemon sensors would be useful.
      2. It was pointed out that there is a need for a central way of publishing lemon configuration in CDB so that it can be used by anyone. Steve Traylen suggested using a PAN function. This is not seen as a good enough option by all VOC's. It would be better to have some support to do this. It was pointed out that lemon ID's do not change so much, in most cases it's a matter of changing the metric.
   3. Flavia rises the need to parametrise the actuator, so that it can be then re-used by many VO's. This way each VOC will not be forced to have 1 ID per type of monitoring. Lemon support team reminds that one can reserve a few hundred ID (CMS is already doing this). By just reserving a bunch of ID's one will not need to change the way Lemon works.
   4. It was also noticed that VOCs have replicated many of the existing sensors/actuators. This could be avoided if VOCs cluster together and share the work.
   5. It is felt that it is better to have good examples. We also need to collect use-cases where VOCs use lemon sensors/metrics/actuators so to start the "clustering" work.
      1. In particular LHCb stressed that they are in favour of changing some of the metrics already used in order to promote shared solutions that are therefore well supported by CF and PES.
      2. This will be re-asked to all VOC's in a future VOC meeting. Flavia volunteered to collect use-cases by all VOCs in a document that will be discussed in the VOC meeting following the one of the web redirector tutorial.
      3. CMS expressed that they are fine and just need more documentation and better examples. CMS agreed to participate to the collection of use-cases in order to learn more about what is used also by other VOCs.
   6. All VOC's are requested to collect these use cases and then circulate them with the aim of finding what is common and evaluate what would be the best way of share commonalities.

Action List

• Action List

-- RicardoSilva - 29-Nov-2010