07:20:22 Hi emilio. So is there a way to see histograms from https://telemetry.mozilla.org/probe-dictionary/?search=mathml relative to mathml.doc_count? 14:16:15 fredw: hey, sorry, missed your ping this morning 14:18:01 emilio: no problem. I was willing to reply to heycam but I don't seem to have access to the data anymore 14:18:15 It seems i need to get my LDAP issue fixed before that 14:19:14 fredw: apparently there's nothing specific for that in the pre-built dashboard / use-counters... I think you can infer it from 14:19:14 https://telemetry.mozilla.org/new-pipeline/evo.html#!aggregates=Median!Mean!5th%2520percentile!25th%2520percentile!75th%2520percentile!95th%2520percentile&cumulative=0&end_date=2019-08-25&include_spill=0&keys=!__none__!__none__&max_channel_version=nightly%252F70&measure=SCALARS_MATHML.DOC_COUNT&min_channel_version=nightly%252F67&processType=*&product=Firefox&sanitize=1&sort_keys=submissions&start_date=2019-07-08&trim=1&use_submis 14:19:14 sion_date=0 and the use-counters page http://georgf.github.io/usecounters/ 14:19:25 fredw: but yeah you need LDAP for the first I think 14:20:01 maybe if I get accesss to the SQL stuff you mentioned, I can write my own thing 14:20:07 fredw: From sql.telemetry.mozilla.org you can build an arbitrary sql query, yeah 14:20:30 * emilio was poking a bit, but it seems like tables are not as documented as one would like... 14:20:36 Also my SQL-foo sucks 14:26:09 emilio: I actually have some difficulty to understand the first graph... but the daily mathml doc count seems to be around on 1 or thousands 14:27:12 fredw: yeah, the first graph is a bunch of medians / etc, which I don't think are particularly relevant 14:27:30 fredw: the graph you care about I think is "submissions" or "sample counts" 14:27:53 fredw: which seem to be between the 1/3k per day 14:28:33 it seems it's lowering down since I commented to mathml people that they can share the data :-/ 14:29:36 fredw: all weekends counters go down, because people go out and that kind of stuff I guess :) 14:31:09 1/3k still seems small population to me, not sure we will really be able to detect any relevant amount for the features we want to disable (well we can say 0% => let's remove it...) 14:32:04 fredw: yeah, note that those counters are for nightly though 14:32:21 fredw: the beta ones are in the 50-90k daily 14:35:18 fredw: and I don't know whether we can look at release data / whether there is release data... Seems to me there should be somewhere 14:36:03 emilio: that would be more useful I guess... also currently we don't warn/count in nightly since the features are disabled anyway... 14:36:11 fredw: right, true that 15:00:00 fredw: for the record, this is what somebody from the telemetry team came up with: 15:00:13 https://www.irccloud.com/pastebin/wmqXNzdg/query 15:01:22 emilio: OK thanks. Normally, you should have tried to say that you wrote this SQL yourself :-) 15:03:42 fredw: well, I did try, I just gave up after a while, no way I could've figured out all those bits myself in a reasonable timeframe :) 15:03:53 LOL 17:07:43 emilio: ok I have recovered by ldap access now. sql.telemetry still seems unaccessible though