WLCG SLS Tape Metrics Review: Proposal Specification

Specification for the reviewed WLCG SLS Tape metrics (see here for background information).

Table of Metrics

metric ID description type unit reporting time window
availability service availability (as defined by each site) mandatory 0-100 since last update (1h)
readfsize_VOx DATA READ: average file size in MB mandatory MB last 24h
writefsize_VOx DATA WRITE: average file size in MB mandatory MB last 24h
readspeed_VOx DATA READ: data transfer rate in MB/s mandatory MB/s last 24h
writespeed_VOx DATA WRITE: data transfer rate in MB/s mandatory MB/s last 24h
datastored_VOx DATA REPOSITORY: total data stored in TB (experiment perspective) mandatory TB last 24h
readwait_VOx TAPE QUEUES: average wait for read in secs optional seconds last 24h
writewait_VOx TAPE QUEUES: average wait for write in secs optional seconds last 24h
dataactive_VOx DATA REPOSITORY: percentage of active data (read over last 12 months) optional 0-100 last 24h

The reporting time window should be understood as a sliding window.

SLS XML update template file

A template SLS XML update file can be found here: WLCG-TAPE-SLS-REVIEWED.pdf

The parts in red color are the ones to be filled in.

Clarifications on some tags:

<availabilitythresholds>
<threshold level="available">75</threshold>
<threshold level="affected">45</threshold>
<threshold level="degraded">15</threshold>
</availabilitythresholds>
It is not recommended to modify those values: just remember them when you calculate the availability of your service. It will be green between 100 and 76%, orange till 45, yellow till 15, then red.

<webpage> http://xxxxxxx </webpage>
(Optional tag) write here the contact email (if you want, or just remove the line) and the webpage where you're reporting more detailed information

<refreshperiod>PT10M</refreshperiod>
Do not touch: this means how often the SLS server will try to read your update xml (10 minutes).

<validityduration>PT2H</validityduration>
This value represents the validity of your data. example: you update the xml every 6hours, this means that that valued should not be smaller than 6hours, otherwise the service will appear gray, as the information is too old. if the information of your tape system are updated every hour, we suggest to put a value of 2H (PT2H) the data structure, all these data will be stored into the SLS DB.

<timestamp>2011-06-07T15:50:05</timestamp>
Format date and time as shown above. Time zone must be Geneva time zone, and it is mandatory.

Prototype implementation - CERN

A prototype implementation of the above SLS tape metrics can be found here:

http://sls.cern.ch/sls/service.php?id=DEV_WLCG_CERN_Tape_Metrics

All metrics are implemented, with the exception of dataactive (percentage of active data) which still needs to be done.

FAQ

  • How is the availability to be computed?
    • This is left over to each site to decide if their tape system is available or not and how much.

  • What are the thresholds to be used in the availabilitythresholds tag?
    • Please use the thresholds proposed above: 75 for available, 45 for affected, 15 for degraded.

  • In addition to the per-VO metrics, is it possible to report a total figure for all VO's?
    • Yes (optionally), by defining a new VO where "VOx" is equal to "ALL". However, only LHC VO's should be included in the total figures.

  • Can we include other metrics which are not included in the above table?
    • For clarity and clearness reasons, no other metrics should be reported.

  • What to do when the optional metrics (average wait, percentage of active data) are not implemented?
    • Simply do not report them. This is clearer than reporting "-1" or similar.

-- GermanCancio - 09-Jun-2011

Topic attachments
I Attachment History Action Size Date Who Comment
PDFpdf WLCG-TAPE-SLS-REVIEWED.pdf r1 manage 48.4 K 2011-06-10 - 10:41 GermanCancio WLCG SLS tape metrics template
Edit | Attach | Watch | Print version | History: r8 < r7 < r6 < r5 < r4 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r8 - 2014-11-20 - TWikiAdminUser
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    HEPTape All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2020 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback