[00:03:24] <groceryheist>	 @chrisalbon cool. how many requests can it handle?
[00:03:24] <AsimovBot>	 04Error: Command “chrisalbon” not recognized. Please review and correct what you’ve written.
[00:04:36] <groceryheist>	 I have a pretty large-scale project in mind tracking changes in article quality over time
[00:05:17] <chrisalbon>	 OKAPI was hitting it for a million a day, but I am not sure it was completely happy with that.
[00:05:55] <groceryheist>	 chrisalbon: great that's very helpful
[00:05:57] <chrisalbon>	 Want to try 3/second and I can let you know if anything goes wrong?
[00:06:25] <chrisalbon>	 Smoke testing prod!
[00:06:57] <groceryheist>	 this is kind of a side project
[00:07:15] <groceryheist>	 i'll do a smaller sample first at 3/sec and let you know when I'm doing it
[00:07:23] <chrisalbon>	 cool
[00:07:27] <groceryheist>	 it might not happen for a couple of weeks
[00:07:55] <groceryheist>	 i would do the smaller sample on the api in any case
[00:08:03] <chrisalbon>	 Since we are all about transparency, you can watch the load yourself https://grafana.wikimedia.org/d/HIRrxQ6mk/ores?orgId=1&refresh=1m
[00:09:07] <groceryheist>	 that's cool
[00:09:30] <chrisalbon>	 https://grafana.wikimedia.org/d/HIRrxQ6mk/ores?viewPanel=13&orgId=1&refresh=1m&from=now-7d&to=now-1m 
[00:09:45] <chrisalbon>	 You can see what happens when it has been overloaded here
[00:09:50] <groceryheist>	 down the road i'm thinking something like 50,000,000 requests
[00:10:01] <chrisalbon>	 like, total or per day
[00:10:04] <groceryheist>	 total
[00:10:14] <chrisalbon>	 cool
[00:11:46] <groceryheist>	 the spike on 10/1 is it getting overloaded?
[00:13:00] <groceryheist>	 i'm not sure what i'm looking at
[00:16:33] <chrisalbon>	 Each one of those spikes is the system overloading to the point that API requests failed
[00:16:54] <chrisalbon>	 So if you hammer the system and see that spike, then you've gone too far :)
[02:05:13] <groceryheist>	  chrisalbon: gotchya thanks
[18:31:40] <wikibugs>	 10ORES, 10Machine Learning Platform (Current): Migrate ORES/Revscoring/etc. repos to Gitlab or Gerrit - https://phabricator.wikimedia.org/T264651 (10calbon)
[20:37:54] <halfak>	 o/ codezee 
[20:37:57] <halfak>	 We meeting?
[23:25:04] <NOSQL>	 halfak: sorry to ping you to my userpage. Figured you might be able to help the editor that was asking me about ORES
[23:28:10] <halfak>	 Oh!  Let me look. 
[23:29:27] <halfak>	 NOSQL: I might have not gotten the ping.  Did you ping Halfak (WMF)? 
[23:29:33] <NOSQL>	 I did
[23:29:43] <NOSQL>	 halfak: FWIW, https://en.wikipedia.org/wiki/User_talk:SQL#AFC-Ores is the link
[23:29:49] <halfak>	 Regretfully that doesn't make it to me anymore since I left the WMF.  "EpochFail" is my volunteer account.  
[23:29:52] * halfak jumps in 
[23:30:38] <NOSQL>	 halfak: Hah, there's a chain of that going on, they asked me because of afc-ores, but I stopped developing tools a while back
[23:30:52] <halfak>	 Ha!  Gotcha.  Nice to hear from you anyway :) 
[23:30:55] <NOSQL>	 I didn't realize you'd left EMF
[23:30:56] <NOSQL>	 *WMF
[23:31:20] <halfak>	 Looks like this is related to the stability issues ORES is having recently.  accraze do you know the status of that?  Has it resolved? 
[23:31:49] <NOSQL>	 Yeah, I'm glacially slow in getting back to messages, I think that one was over a week ago, sorry :(
[23:36:06] <halfak>	 No sweat.  I appreciate you made the connection and found me here.  
[23:36:24] <halfak>	 I just replied NOSQL.  Hopefully I can help out SD0001 or find someone who can :) 
[23:36:32] <NOSQL>	 halfak: tyvm!