« first day (1913 days earlier)      last day (1899 days later) » 

10:00 PM
@ArtOfCode Yeah, cuz now we have better timing data
So graphs of average db/view time over time
 
tpu- feedback received
[ SmokeDetector | MS ] Potentially bad keyword in body, potentially bad keyword in title (2): Help! What is wrong with this Coleus? Overwatered? Underwatered? by M Clarke on gardening.SE
 
@thesecretmaster Oh, gotcha
 
[
  { name: 'series-name', data: [ [x0, y0], [x1, y1], ... ] },
  ...
]
@thesecretmaster ^ that's the format that routes need to return to go into a graph
as JSON
 
[ SmokeDetector | MS ] Bad keyword in body (97): Some help on Bayes Statistics/probability problem by dumbboi on math.SE (@FrenzyLi)
[ SmokeDetector | MS ] Potentially bad keyword in body (1): Hybris 30 days license expired by JuanDroid on stackoverflow.com
 
10:19 PM
@ArtOfCode So... I need to deal with i.e. averaging data per-hour server-side?
 
yup
Not difficult, use group_by_hour_of_day and SQL AVG
look at graphs_controller for examples
 
!!/pull
 
@SmokeDetector n testing SILLY cOmmeNT capitalisation
 
10:23 PM
@angussidney That report was a question; questions cannot be marked as NAAs.
 
@SmokeDetector f testing SILLY cOmmeNT capitalisation
\o/
 
naa feedback received
 
@ArtOfCode Except the data is in redis
 
oh, bugger
well I have approximately zero idea how you do that, then
 
Maybe the graphing lib has magic?
 
10:29 PM
Chartkick doesn't
It's aimed straight at doing it with SQL
Highcharts... we'll see
 
I mean, it's easy to get x,y pairs from redis, I just want to be able to smooth it.
 
Might be able to do it in JS
 
I can also write a redis module to round timestamps to the nearest n, but that seems like overkill.
 
...just a little
 
I think this should happen client side though
Maybe there's a ruby lib for it?
Also, you know that the specific time incriment I'm rounding to doesn't matter, I just want some neat graphical representation of page load speed over time?
 
10:37 PM
[ SmokeDetector | MS ] Link at end of answer (60): Polygons for postcodes in Spain by Alonso V on gis.SE
 
Maybe there's a better way to do this?
 
fp feedback received
 
Okay, I've had a look at doing it in JS... let's not
given how disgusting that code would be, ruby it is
 
@thesecretmaster how'd you get (x, y) pairs from redis?
 
10:44 PM
@ArtOfCode zrange
 
If you can give me an array of pairs, I can probably figure out a decent way to stick a moving average on them in ruby
 
Oh sure, 1 sec
zrevrange request_timings/total/by_action/<controller>#<action> 0 -1 WITHSCORES
@ArtOfCode ^^
 
right
having realised that I'm dumb, JS might also work
 
:P
Um...
Something is broken
 
nope, no, I was right, it won't
 
10:49 PM
AIM is deeply unhappy
 
usually is
 
Oh yes that again
Ssssh nothing happened
Ehm... something happened
Okey, 1 sec
@Undo Around?
Alrighty, purging a lot of requests
 
iBug/EC2 received failover signal.
fp feedback received
Restart: API quota is 19960.
fp feedback received
 
MS back alive, sorry y'all
 
[ SmokeDetector | MS ] Potentially bad keyword in body (1): Where can I find an aviation headset to mini-jack adapter? ✏️ by user9943887 on aviation.SE (@TetsuyaYamamoto)
[ SmokeDetector | MS ] Potentially bad keyword in body, potentially bad keyword in title (2): Help! What is wrong with this Coleus? Overwatered? Underwatered? by M Clarke on gardening.SE
[ SmokeDetector | MS ] Blacklisted website in answer (92): Why won't my files import into garageband? ✏️ by New Beginningboy on apple.SE
[ SmokeDetector | MS ] Blacklisted website in answer (92): Why won't my files import into garageband? ✏️ by New Beginningboy on apple.SE
 
11:06 PM
def moving_avg(pts)
  yv = pts.map(&:last)
  ad = yv.map.with_index { |v, i| yv[([0, i-3].max)..([i+3, yv.size].min)].sum / (([0, i-3].max)..([i+3, yv.size].min)).size.to_f }
  pts.map.with_index { |e, i| [e[0], ad[i]] }
end
@thesecretmaster ^
 
Will it be expensive?
 
Shouldn't be, but lemme benchmark it
What kinda size of dataset am I expecting?
 
What's action/controller on status-update.json?
 
status#status_update
 
[ SmokeDetector | MS ] Body starts with title and ends in url (98): compare 2 stl files by Marcin on blender.SE
 
11:10 PM
About 100k, if you want full history. Stopping it at about a month gives us a couple thousand I think.
 
[ SmokeDetector | MS ] Body starts with title and ends in url (98): compare 2 stl files by Marcin on blender.SE
 
@thesecretmaster 100k pairs?
 
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): How fix AndroidManifest.xml ✏️ by Dana Dahlstrom on stackoverflow.com
 
okey doke
 
11:11 PM
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): How fix AndroidManifest.xml ✏️ by Dana Dahlstrom on stackoverflow.com
 
1 for each time the route gets hit
 
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): Android Studio Auto Generated public final class R file ✏️ by Dana Dahlstrom on stackoverflow.com
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): Android Studio Auto Generated public final class R file ✏️ by Dana Dahlstrom on stackoverflow.com
 
!!/location
 
@EriktheOutgolfer teward/Lunar Eclipse
@EriktheOutgolfer iBug/EC2
 
!!/standby ec2
 
@thesecretmaster Still need me? 40% here, 60% fighting 3D printer
 
Nah, thanks for checking though
 
k, thanks for handling whatever it was
 
fp feedback received on compare 2 stl files [MS]
 
11:13 PM
@Undo Well, I asked a SO question...
 
0
Q: After loading from RDB dump, no keys have TTL set

thesecretmasterI have a redis database which is configured to persist with RDB. It grows very quickly, but has maxmemory 80mb and maxmemory-policy volatile-ttl. The reason I need volatile ttl is because there are certain keys which absolutely cannot be evicted without breaking the application. The vast majority...

 
@thesecretmaster 70ms
 
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): How long does caulk need to dry in a 1" gap? by Answer the question on diy.SE
 
11:15 PM
I set it to benchmark 10 times to average it and fully expected it to take a few seconds for a million pairs... nope, under a second
 
[ SmokeDetector | MS ] Bad keyword with email in answer, pattern-matching email in answer, potentially bad keyword in answer (185): MySQL sum amounts for transactions related to loans, in reverse order by Norisa Waye on stackoverflow.com
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): After Windows crashed custom classes converted into the R class by Dana Dahlstrom on stackoverflow.com
 
to be fair, it is only O(n)
 
I won't use it unless it runs in O(1) :P
 
fp feedback received on How fix AndroidManifest.xml [MS]
 
Good luck with that one ;)
 
@thesecretmaster yeeeeah... soooo, this is how a friend of a cousin of an uncle of mine got fired... he shouted at everything above O(1)... obviously, his coworkers weren't too happy... ;)
 
fp feedback received
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): java and xml files corrupted by weird XML in Android Studio(3.2.1) by Dana Dahlstrom on stackoverflow.com
 
GOSH DARN TYPOS
 
fp feedback received
[ SmokeDetector | MS ] Bad keyword with email in answer (85): Issue symfony 4.2 and postgresql by Emmanuel Beranger on stackoverflow.com
 
Welp, now I need to purge every single request from our logs
That's fun
 
11:27 PM
naa feedback received
[ metasmoke-deploy ] deploy started by thesecretmaster
[ metasmoke ] ci/circleci: js failure on e01eb94: Your tests failed on CircleCI
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): My Activity Class turn into an xml file by Dana Dahlstrom on stackoverflow.com
 
@ArtOfCode Any objection s to me purging all the redis logs ever?
 
[ metasmoke ] ci/circleci: rubocop failure on e01eb94: Your tests failed on CircleCI
 
nope
 
Say bye :'(
 
11:38 PM
@SmokeDetector 7 answers saying the same thing is way too many. I'm calling it TP, as we'd would want to block this.
 
tpu- feedback received
[ metasmoke ] ci/circleci: rails-tests failure on e01eb94: Your tests failed on CircleCI
tpu- feedback received on How fix AndroidManifest.xml [MS]
 
What is status_update.json?
It gets hit so frequently
 

« first day (1913 days earlier)      last day (1899 days later) »