Alert.png The wiki is deprecated and due to be decommissioned by the end of September 2022.
The content is being migrated to other supports, new updates will be ignored and lost.
If needed you can get in touch with EGI SDIS team using operations @ egi.eu.

Difference between revisions of "Service Level Target - ROD performance index"

From EGIWiki
Jump to navigation Jump to search
Line 38: Line 38:


#When an Operations Centre get a ticket from COD about ROD performance, the Operations Centre should provide GGUS ticket or a link to SAM or the oeprations dashboard downtime page in GOC DB.  
#When an Operations Centre get a ticket from COD about ROD performance, the Operations Centre should provide GGUS ticket or a link to SAM or the oeprations dashboard downtime page in GOC DB.  
## if
#Based on GGUS trouble tickets referenced in prerequisites, or on the GGUS ticket opened by the Operations Centre to MyEGI requesting for A/R recualculation, or GOC DB service downtime entry, COD, knowing when the problem occurred,<br>can remove the metrics items for given days from final report pdf.
#Based on GGUS trouble tickets referenced in prerequisites, or on the GGUS ticket opened by the Operations Centre to MyEGI requesting for A/R recualculation, or GOC DB service downtime entry, COD, knowing when the problem occurred,<br>can remove the metrics items for given days from final report pdf.



Revision as of 12:22, 25 January 2012

Main EGI.eu operations services Support Documentation Tools Activities Performance Technology Catch-all Services Resource Allocation Security



The ROD performance index (formerly known as ROD OLA metric) was introcuded to track the level of Grid Oversight service delivered by Operations Centres according to Resource Provider OLA.

The index was accepted during Technical Forum 2011 in Lyon and is available on EGI Operations Portal.

Definition

ROD performance index is calculated monthly from the data gathered by EGI Operations Portal. It does not take into account weekends.

ROD performance index is a sum of:

  • No. of ticket expired* appearances in operations dashboard daily
  • No. alarms older than 72h appearances in operations dashboard daily

The threshold was set to 10 items. Above this value ROD teams has to provide explanation and provide a plan of improbement of the oversight service.


expired ticket = a ticket inOperations Portal dashboard with "Expiration date" set at the past. The "Expiration date" field is set according to escalation procedure, but can be freely changed by ROD. It refers to the date when the status of issue should be checked next time.

Performance reports

Performance reports can be found on Availability and reliability monthly statistics wiki page

Recalculation procedure in case of intervention on the NGI SAM or the operations dashboard

Prerequisite:

  1. In case of problems with synchronization between regional operations dashboard the Resource infrastructure Provider should create a GGUS ticket to Operations Portal team.
  2. In case of problems with the regional SAM, Resource infrastructure Provider should create a GGUS ticket to SAM team.
  3. In case of work carried out on regional SAM or operations dashboard the Resource infrastructure Provider should declare downtime in GOC DB.

Procedure steps:

  1. When an Operations Centre get a ticket from COD about ROD performance, the Operations Centre should provide GGUS ticket or a link to SAM or the oeprations dashboard downtime page in GOC DB.
    1. if
  2. Based on GGUS trouble tickets referenced in prerequisites, or on the GGUS ticket opened by the Operations Centre to MyEGI requesting for A/R recualculation, or GOC DB service downtime entry, COD, knowing when the problem occurred,
    can remove the metrics items for given days from final report pdf.

Future plans

In the future the metric will also include no. alarms closed in NON-OK status without explanation. This will need some implementation effort.

Issues to be implemented:

  • Taking into account holidays periods in alarms ageing
  • Automatic check if site/node is in downtime while alarm is closing
  • Automatic check if node is not in production while alarm is closing
  • In case of SCHEDULED interventions, the monthly metrics calculation should automatically take the scheduled downtime into account. At the time the metrics are computed, the application which does such calculation should access the GOC PI to determine which regional nagios machines were in downtime, and include that restriction in the calculation.