USAG meeting - Theme: Specification of Direct Site Notification for ALL GGUS tickets. New TPM role.

Europe/Zurich
28-R--006 (CERN)

28-R--006

CERN

Description
User Support Advisory Group (USAG) meeting This group is, in EGEE III, the successor of the EGEE II Executive Support Committe (ESC).

Information twiki

Dial-in numbers: +41227676000 (English, Main)
Access codes: 0177773 (Leader)
0187746 (Participant)
Leader site: https://audioconf.cern.ch/call/0177773
Participant site: https://audioconf.cern.ch/call/0187746

  1. Agenda approval
  2. Comments on minutes from the October 30th USAG meeting [here]
  3. Continuous user support assessment as required by MSA-1.6 via the weekly escalation reports
  4. Comments on the QR2 conclusions for SA1 management:
    Extract from Maite's Nov. 18th email:
    (...) 
    3. N of user support tickets created: the same, big increase in the last
    quarter; it would be worth to understand what are the support units that
    caused this and understand the reason (if any).
    4. User support Tickets solution time: there is a peak in September
    worth to investigate or correlate with the number of tickets. 
    
  5. News from the Direct Routing of GGUS tickets to Sites following the Nov. 12th GDB [see GDB slides] and the Nov. 14th WLCG workshop [see WLCG workshop slides] presentations.
  6. Regional helpdesks and other requests for interfacing to GGUS.
  7. Grid middleware development - savannah - GGUS interface.
  8. Review Action List [go to the end of these notes]. Please see here whether action 20080506-1 can be closed.
  9. Next meeting date: Thu 15 Jan 2009 @ 9:30 CET
  10. A.O.B.
Last Update: 2008-11-27 by M.Dimou
draft_for_agenda_point_6
USAG meeting 27th November 2008


Click on the agenda
to follow links to the documents discussed at the meeting!



Attendance
----------
GGUS - Torsten Antoni, Helmut Dres, Guenter Grein
CERN - Diana Bosio, Maria Dimou (chairperson)
CE   - Jan Kmunicek
UK/I - John Gordon, Claire Devereux
IT   - Riccardo Brunetti

Apologies:
----------

Notes by Diana

Agenda:
-------
1. Agenda approval
2. Comments on minutes from the October 30th USAG meeting
3. Continuous user support assessment as required by MSA-1.6 via the weekly escalation reports
4. Comments on the QR2 conclusions for SA1 management:

      Extract from Maite's Nov. 18th email:
      (...)
      3. N of user support tickets created: the same, big increase in the last
      quarter; it would be worth to understand what are the support units that
      caused this and understand the reason (if any).
      4. User support Tickets solution time: there is a peak in September
      worth to investigate or correlate with the number of tickets.

5. News from the Direct Routing of GGUS tickets to Sites following the
   Nov. 12th GDB [see agenda] and the Nov. 14th WLCG workshop [see
   agenda] presentations.

          * Decision on suspended/test/uncertified sites.
          * Extended TPM role:
                o first line support for JRA1 JRA1 All-Hands' presentation and on-going GGUS-savannah interface development.
                o first escalation steps for tickets . Comment on the GGUS ticket escalation procedure.
                o revision of documentation
6. Regional helpdesks and other requests for interfacing to GGUS.
7. Grid middleware development - savannah - GGUS interface.
8. Review Action List [go to the end of these notes]. Please see here whether action 20080506-1 can be closed.
9. Next meeting date: Thu 15 Jan 2009 @ 9:30 CET
10. A.O.B.


Discussion
----------


1. Agenda was approved
2. Comments on minutes from the October 30th USAG meeting
No comments. Minutes were approved.
3. Continuous user support assessment as required by MSA-1.6 via the weekly escalation reports

As an aside, the units regarding operational tools will be deleted
from the EMT reports as they are already present in the ROC reports.

The escalation reports were examined and it was concluded that the
most problematic units are those in the EMT reports. It is probably
still a legacy of the problems that JRA1 had in staffing the software
units. Nonetheless Diana and Maria will take up the reports in between
meetings with the units concerned: Maria for anything related to VO
support and VOMS, Diana for SA3 and the other software units.

John asked about the metrics of the quality group. Diana said there is
a website. Here is the link:
http://project-egee-iii-na1-qa.web.cern.ch/project-EGEE-III-NA1-QA/EGEE-III/QoS/QoS.htm

Maria also reported how the average assignment time for the TPM is
going up again: last week about 50% of the tickets only were assigned
under 1 hour. To be monitored.

4. Comments on the QR2 conclusions for SA1 management:

Point 3 in Maite's e-mail (see agenda) is due mainly to an increase in
monitoring activities that use GGUS. For point 4 it would be worth
differentiating by ticket category (incident, change request,
documentation).

Hereafter find Maria's e-mail to Maite with the USAG response to the
comments and Maite's answer.

====Maria's e-mail==========================
  
Dear Maite, points 3 and 4 were discussed at the USAG meeting today
(see agenda item 4 in
http://indico.cern.ch/conferenceDisplay.py?confId=45916 ). Here are
our comments:

On point 3:
-----------
The increase of GGUS tickets is a result of GGUS' increasing
popularity and improving efficiency. Via the direct routing to sites
since July 3rd 2008, the LHC experiment VOs started to trust GGUS much
more and use it for reporting and monitoring. All WLCG and EGEE
Operations' meetings discuss in our days based on GGUS ticket numbers.

On point 4:
-----------

More detailed analysis by the GGUS developer H.Dres showed that there
was a drop in the number of GGUS tickets in August, certainly due to
holidays. This created th impression of a 'big increase' in September
2008. In fact, there were even more tickets in October. A suggestion
by the CERN ROC manager D.Bosio for the next round of metrics was to
distinguish per Ticket Type, i.e. Incident, Change request, Spam,
Documentation. The TPMs are reminded to use correctly, even modify the
value of the 'Ticket Type' field, as it will signify different
required solution times.

TPM and supporters should be trained to use this field correctly and
change the category of a ticket if necessary, sice most people
nowadays do not use this field correctly.

===Maite's reply=============================================

Thanks a lot Maria and the GGUS and USAG teams for this detailed
analysis. I'll include the conclusions in next QR.  I am very happy to
know that the conclusion is an increase of the confidence and use of
GGUS.  Congratulations for the good work.  Thanks a lot,

Maite


5. News from the Direct Notification of GGUS tickets to Sites

     * Decision on suspended/test/uncertified sites.

Of the above, only uncertified ites will be included, as sometimes
they already have a support structure in place. If this proves to
clutter the list of site with no additional benefit they will be
removed.

     * Extended TPM role:
         o first line support for JRA1 JRA1 All-Hands' presentation and on-going GGUS-savannah interface development.
         o first escalation steps for tickets . Comment on the GGUS ticket escalation procedure.
         o revision of documentation

TPMs should be told and trained of this shift of focus of their role
and the additional responsibility of being the second escalation
step.

TPM training should be revived, especially when new teams replace the
odl ones.Torsten will include this point in the presentation at the
next face to face SA1 coordination management meeting.

6. Regional helpdesks and other requests for interfacing to GGUS.

See Guenter's document attached to the agenda. It is a proposal to
reduce and limit the number of possible interfaces to GGUS. The most
recommended way to interface to GGUS currently is via web services.

7. Grid middleware development - savannah - GGUS interface.
The savannah interface was described while talking of escalation
reports. It will go live in January 2009.


8. Review Action List [go to the end of these notes]

9. Next meeting date: Thu 15 Jan 2009 @ 9:30 CET
10. A.O.B.
Two points for next meeting agenda:

A. The SLA/OLA for the various SU will be discussed at the next USAG meeting.
B. Feedback to the prototype for direct notification to sites.The
prototype will be available on the training system in December.


ACTION LIST
-----------
20080506-1:
(Was: #164.14) Provide information on the CIC VO XML dump
Cyril
last message from Gilles explaining the situation:
http://indico.cern.ch/getFile.py/access?resId=0&materialId=3&confId=37954
Carry forward until autumn
Development work is now advanced on the CIC portal. This action can be completed soon.
Gilles will send a reminder to Cyril. Maria also sent one on 2008-11-04.

Done

200806-6:
Report progress on the definition of escalation authority for software
tickets.
Diana and Maria
report: requirement from Cal. conclusion was that button should not be there,
but tickets should be escalated to the usag.
Maria: we should find a way to tell the user. Re-opened https://savannah.cern.ch/support/index.php?104440
Item is now due for the January release
Done


200807-6:
revive the thread of work on metrics and associated reports
Diana and Guenter
will be moved to savannah for a weekly reminder and check of progress
https://savannah.cern.ch/support/?106366
Done

20081127-1
Monitor and report statistics on assignment time of tickets by TPM
Diana
Pending

20081127-2
Ask SW documentation on their web service interface to GGUS, and see
if it is possible to reuse their code for other ROCs using the same
ticketing system (1or0).
Guenter
Done

There are minutes attached to this event. Show them.
The agenda of this meeting is empty