Michael added a comment.

  The patch above (Ib7bfacf88 
<https://gerrit.wikimedia.org/r/#/q/Ib7bfacf88eef436ed0f34a16e9d39eb0c053a44f>/02109a44
 
<https://gerrit.wikimedia.org/g/operations/puppet/+/02109a444a6ece2f480aee183dc02074d0f3734d>)
 seemingly resolved the original issue, but some new errors started appearing 
(though much less common than before).
  
  Those new errors seem to come from a timeout in Termbox/Kubernetes which 
result in a timeout in the php/mediawiki code.
  
  The following two links illustrate that error for the rendering process of a 
single item: 庞家村 (Q14836111) <https://www.wikidata.org/wiki/Q14836111>
  
  Termbox/k8s: 
https://logstash.wikimedia.org/app/kibana#/doc/logstash-*/logstash-syslog-2020.06.15/syslog?id=AXK3_osLZmYAikdJbyT-&_g=h@e3739c2
  Mediawiki: 
https://logstash.wikimedia.org/app/kibana#/doc/logstash-*/logstash-mediawiki-2020.06.15/mediawiki?id=AXK3_omyK6TSR36GrxZd&_g=h@e1c60c6
  
  This was judged to be not pressing (@Tarrow .@Addshore  please correct me if 
I'm talking rubbish here). Currently, the internal timeouts seem to be 3 
seconds both for the Mediawiki side (ssrServerTimeout 
<https://github.com/wikimedia/mediawiki-extensions-Wikibase/blob/57e673d69ee8357a65a4746ed5dd289fc9f6b61b/docs/topics/options.md#ssrservertimeout>?)
 and for the Termbox side. It is not clear what we can do about those errors. 
While the internal requests **should** not take longer than 3 seconds, they 
sometimes do.
  
  **Follow-ups:**
  
  - It would be good if a timeout noticed and reported by Termbox would be 
treated differently by the Mediawiki error logging infrastructure than a 
timeout of Termbox/k8s itself: T255436: Termbox Error Logging Should 
Differentiate between RemoteRenderer and Service timeouts. 
<https://phabricator.wikimedia.org/T255436>.
  - there is an metric counting such request errors, but currently it is not 
tracked: T255426: Create Grafana tracking 
`wikibase.view.TermboxRemoteRenderer.requestError` and maybe add alert 
<https://phabricator.wikimedia.org/T255426>
  - we currently do not have dashboard tracking the errors in Termbox 
SSR-service/Kibana: T255437: Create Logstash Dashboard Tracking Termbox Errors 
<https://phabricator.wikimedia.org/T255437>

TASK DETAIL
  https://phabricator.wikimedia.org/T255410

EMAIL PREFERENCES
  https://phabricator.wikimedia.org/settings/panel/emailpreferences/

To: Michael
Cc: JMeybohm, WMDE-leszek, Pablo-WMDE, Tarrow, Jakob_WMDE, Addshore, Aklapper, 
Michael, darthmon_wmde, Nandana, Lahi, Gq86, GoranSMilovanovic, QZanden, 
LawExplorer, _jensen, rosalieper, Scott_WUaS, Wikidata-bugs, aude, 
Lydia_Pintscher, Mbch331
_______________________________________________
Wikidata-bugs mailing list
Wikidata-bugs@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs

Reply via email to