Ladsgroup added a comment.

  In T246898#6017915 <https://phabricator.wikimedia.org/T246898#6017915>, 
@Addshore wrote:
  
  > So, these still happen, but at a much lower rate.
  > This is probably still something that we would like to iron out and that we 
would probably tackle with some locks for updating entities higher up in the 
application (not in sql).
  
  Sorry if I'm missing something obvious but I don't see it's happening 
anymore. I have looked at the numbers again and the differed update failing for 
wikidata in the past 24 hours is only six cases which none is related to term 
store (one is echo not working, several others is mediawiki not being able to 
push jobs and fataling: "Could not enqueue jobs: Unable to deliver all events: 
503: Service Unavailable") but error like this is gone. The value in the graphs 
I posted is not down to zero because it's graph of all errors in mediawiki for 
wikidatawiki and not this particular error. Even at failure rate of six per 
day, which is 6 per 712K ~= 1 / 100K = 10 ppm or 0.0001%. I think that's pretty 
normal, edits fail to save at almost the same rate (5 every 24h)

TASK DETAIL
  https://phabricator.wikimedia.org/T246898

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

To: Ladsgroup
Cc: Ladsgroup, brennen, Magnus, Agusbou2015, Addshore, Liuxinyu970226, 
Aklapper, Hazizibinmahdi, darthmon_wmde, Nandana, Lahi, Gq86, 
GoranSMilovanovic, QZanden, LawExplorer, _jensen, rosalieper, Scott_WUaS, 
Jonas, Wikidata-bugs, aude, Lydia_Pintscher, Jdforrester-WMF, Mbch331, Rxy, 
Jay8g, Krenair
_______________________________________________
Wikidata-bugs mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs

Reply via email to