As a product owner, I want to avoid logging events at a high sampling threshold for a long period, so that my queries run faster and we use less disk space.
This should be deployed once both the new language switcher button affordance and language switcher overlay have been in production for a couple weeks.
| Subject | Repo | Branch | Lines +/- | |
|---|---|---|---|---|
| Reduce sampling rate for language switcher | operations/mediawiki-config | master | +1 -1 |
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Duplicate | • bmansurov | T127212 Remove sampling rate artifacts for language switcher after enough data collected | |||
| Declined | dr0ptp4kt | T128214 Issues with A/B test bucketing for language selector | |||
| Resolved | None | T128917 Head-to-head A/B test language switcher alternative |
Change 272724 had a related patch set uploaded (by Bmansurov):
Reduce sampling rate for language switcher
Given there are issues with bucketing I suspect we'll need to bump this to the sprint after this one?
As a formality does this need to go back to To Do so it can be SWAT'd? I'm okay with it being SWAT'd. At this point there's plenty of data.
Let's hold off on this piece right now. The data are presently inconclusive, and part of this has to do with the relatively lower number of events with version B, structured-overlay. I'll open a fresh task to make the A/B test head-to-head!
Overlay start of funnel select left(t2.timestamp, 8) ts, t2.event_languageOverlayVersion, t1.event_languageButtonTappedBucket, count(*) from MobileWebLanguageSwitcher_15302503 t2 inner join MobileWebLanguageSwitcher_15302503 t1 on t2.event_FunnelToken = t1.event_FunnelToken inner join MobileWebLanguageSwitcher_15302503 t0 on t1.event_FunnelToken = t0.event_FunnelToken where t2.timestamp > '20160227' and t2.timestamp < '20160304' and t2.event_event = 'languageListLoaded' and t2.event_MobileMode = 'stable' and t1.timestamp > '20160227' and t1.timestamp < '20160304' and t1.event_event = 'languageButtonTap' and t0.timestamp > '20160227' and t0.timestamp < '20160304' and t0.event_event = 'pageLoaded' and t0.event_beaconCapable = 1 group by ts, t2.event_languageOverlayVersion, t1.event_languageButtonTappedBucket; ts event_languageOverlayVersion event_languageButtonTappedBucket count(*) 20160227 simpler-overlay 1-4 taps 69 20160227 simpler-overlay 20+ taps 12 20160227 simpler-overlay 5-20 taps 41 20160227 simpler-overlay unknown 145 20160227 structured-overlay 1-4 taps 9 20160227 structured-overlay 20+ taps 1 20160227 structured-overlay 5-20 taps 2 20160227 structured-overlay unknown 20 20160228 simpler-overlay 1-4 taps 94 20160228 simpler-overlay 20+ taps 10 20160228 simpler-overlay 5-20 taps 26 20160228 simpler-overlay unknown 162 20160228 structured-overlay 1-4 taps 7 20160228 structured-overlay 20+ taps 1 20160228 structured-overlay 5-20 taps 1 20160228 structured-overlay unknown 22 20160229 simpler-overlay 1-4 taps 80 20160229 simpler-overlay 20+ taps 13 20160229 simpler-overlay 5-20 taps 32 20160229 simpler-overlay unknown 166 20160229 structured-overlay 1-4 taps 5 20160229 structured-overlay 20+ taps 2 20160229 structured-overlay 5-20 taps 2 20160229 structured-overlay unknown 9 20160301 simpler-overlay 0 taps 1 20160301 simpler-overlay 1-4 taps 84 20160301 simpler-overlay 20+ taps 8 20160301 simpler-overlay 5-20 taps 26 20160301 simpler-overlay unknown 142 20160301 structured-overlay 1-4 taps 8 20160301 structured-overlay 20+ taps 5 20160301 structured-overlay 5-20 taps 5 20160301 structured-overlay unknown 13 20160302 simpler-overlay 0 taps 3 20160302 simpler-overlay 1-4 taps 69 20160302 simpler-overlay 20+ taps 12 20160302 simpler-overlay 5-20 taps 27 20160302 simpler-overlay unknown 137 20160302 structured-overlay 1-4 taps 7 20160302 structured-overlay 20+ taps 1 20160302 structured-overlay 5-20 taps 2 20160302 structured-overlay unknown 11 20160303 simpler-overlay 0 taps 25 20160303 simpler-overlay 1-4 taps 73 20160303 simpler-overlay 20+ taps 7 20160303 simpler-overlay 5-20 taps 24 20160303 simpler-overlay unknown 117 20160303 structured-overlay 0 taps 4 20160303 structured-overlay 1-4 taps 4 20160303 structured-overlay 20+ taps 1 20160303 structured-overlay 5-20 taps 7 20160303 structured-overlay unknown 9 Completed funnels select left(t3.timestamp, 8) ts, t2.event_languageOverlayVersion, t1.event_languageButtonTappedBucket, t3.event_exitModal, t3.event_searchInputHasQuery, count(*) from MobileWebLanguageSwitcher_15302503 t3 inner join MobileWebLanguageSwitcher_15302503 t2 on t3.event_FunnelToken = t2.event_FunnelToken inner join MobileWebLanguageSwitcher_15302503 t1 on t2.event_FunnelToken = t1.event_FunnelToken inner join MobileWebLanguageSwitcher_15302503 t0 on t1.event_FunnelToken = t0.event_FunnelToken where t3.timestamp > '20160227' and t3.timestamp < '20160304' and t3.event_event = 'exitModal' and t3.event_mobileMode = 'stable' and t2.timestamp > '20160227' and t2.timestamp < '20160304' and t2.event_event = 'languageListLoaded' and t1.timestamp > '20160227' and t1.timestamp < '20160304' and t1.event_event = 'languageButtonTap' and t0.timestamp > '20160227' and t0.timestamp < '20160304' and t0.event_event = 'pageLoaded' and t0.event_beaconCapable = 1 group by ts, t2.event_languageOverlayVersion, t1.event_languageButtonTappedBucket, t3.event_exitModal, t3.event_searchInputHasQuery; ts event_languageOverlayVersion event_languageButtonTappedBucket event_exitModal event_searchInputHasQuery count(*) 20160227 simpler-overlay 1-4 taps dismissed 0 1 20160227 simpler-overlay 1-4 taps tapped-on-result 0 60 20160227 simpler-overlay 20+ taps dismissed 0 1 20160227 simpler-overlay 20+ taps tapped-on-result 0 6 20160227 simpler-overlay 5-20 taps dismissed 0 1 20160227 simpler-overlay 5-20 taps tapped-on-result 0 34 20160227 simpler-overlay unknown dismissed 0 1 20160227 simpler-overlay unknown dismissed 1 2 20160227 simpler-overlay unknown tapped-on-result 0 108 20160227 simpler-overlay unknown tapped-on-result 1 6 20160227 structured-overlay 1-4 taps tapped-on-result 0 9 20160227 structured-overlay 20+ taps tapped-on-result 0 1 20160227 structured-overlay 5-20 taps tapped-on-result 0 1 20160227 structured-overlay unknown dismissed 0 1 20160227 structured-overlay unknown tapped-on-result 0 12 20160228 simpler-overlay 1-4 taps dismissed 0 4 20160228 simpler-overlay 1-4 taps tapped-on-result 0 82 20160228 simpler-overlay 20+ taps tapped-on-result 0 12 20160228 simpler-overlay 5-20 taps tapped-on-result 0 24 20160228 simpler-overlay unknown dismissed 0 11 20160228 simpler-overlay unknown dismissed 1 4 20160228 simpler-overlay unknown tapped-on-result 0 87 20160228 structured-overlay 1-4 taps tapped-on-result 0 8 20160228 structured-overlay 20+ taps tapped-on-result 0 1 20160228 structured-overlay unknown dismissed 0 1 20160228 structured-overlay unknown tapped-on-result 0 13 20160229 simpler-overlay 1-4 taps dismissed 0 5 20160229 simpler-overlay 1-4 taps dismissed 1 1 20160229 simpler-overlay 1-4 taps tapped-on-result 0 64 20160229 simpler-overlay 20+ taps tapped-on-result 0 8 20160229 simpler-overlay 5-20 taps dismissed 0 1 20160229 simpler-overlay 5-20 taps tapped-on-result 0 29 20160229 simpler-overlay unknown dismissed 0 9 20160229 simpler-overlay unknown dismissed 1 3 20160229 simpler-overlay unknown tapped-on-result 0 108 20160229 structured-overlay 1-4 taps tapped-on-result 0 3 20160229 structured-overlay 20+ taps tapped-on-result 0 2 20160229 structured-overlay 5-20 taps tapped-on-result 0 2 20160229 structured-overlay unknown dismissed 1 1 20160229 structured-overlay unknown tapped-on-result 0 1 20160301 simpler-overlay 0 taps tapped-on-result 0 6 20160301 simpler-overlay 1-4 taps dismissed 0 3 20160301 simpler-overlay 1-4 taps tapped-on-result 0 59 20160301 simpler-overlay 20+ taps tapped-on-result 0 8 20160301 simpler-overlay 5-20 taps dismissed 0 1 20160301 simpler-overlay 5-20 taps tapped-on-result 0 18 20160301 simpler-overlay unknown dismissed 0 12 20160301 simpler-overlay unknown dismissed 1 3 20160301 simpler-overlay unknown tapped-on-result 0 84 20160301 simpler-overlay unknown tapped-on-result 1 3 20160301 structured-overlay 1-4 taps dismissed 0 1 20160301 structured-overlay 1-4 taps tapped-on-result 0 3 20160301 structured-overlay 20+ taps tapped-on-result 0 3 20160301 structured-overlay 5-20 taps tapped-on-result 0 4 20160301 structured-overlay unknown tapped-on-result 0 10 20160302 simpler-overlay 0 taps tapped-on-result 0 3 20160302 simpler-overlay 1-4 taps dismissed 0 1 20160302 simpler-overlay 1-4 taps tapped-on-result 0 79 20160302 simpler-overlay 20+ taps dismissed 0 1 20160302 simpler-overlay 20+ taps tapped-on-result 0 8 20160302 simpler-overlay 5-20 taps dismissed 0 3 20160302 simpler-overlay 5-20 taps tapped-on-result 0 19 20160302 simpler-overlay unknown dismissed 0 8 20160302 simpler-overlay unknown dismissed 1 5 20160302 simpler-overlay unknown tapped-on-result 0 77 20160302 structured-overlay 1-4 taps tapped-on-result 0 8 20160302 structured-overlay 5-20 taps tapped-on-result 0 1 20160302 structured-overlay unknown tapped-on-result 0 3 20160303 simpler-overlay 0 taps dismissed 0 2 20160303 simpler-overlay 0 taps tapped-on-result 0 24 20160303 simpler-overlay 0 taps tapped-on-result 1 1 20160303 simpler-overlay 1-4 taps dismissed 0 3 20160303 simpler-overlay 1-4 taps tapped-on-result 0 66 20160303 simpler-overlay 20+ taps tapped-on-result 0 5 20160303 simpler-overlay 5-20 taps dismissed 0 2 20160303 simpler-overlay 5-20 taps tapped-on-result 0 23 20160303 simpler-overlay unknown dismissed 0 12 20160303 simpler-overlay unknown dismissed 1 3 20160303 simpler-overlay unknown tapped-on-result 0 75 20160303 structured-overlay 0 taps dismissed 0 1 20160303 structured-overlay 0 taps tapped-on-result 0 2 20160303 structured-overlay 1-4 taps tapped-on-result 0 21 20160303 structured-overlay 20+ taps tapped-on-result 0 1 20160303 structured-overlay 5-20 taps tapped-on-result 0 3 20160303 structured-overlay unknown tapped-on-result 0 6 20160303 structured-overlay unknown tapped-on-result 1 1
@bmansurov I updated the description to indicate we should deploy this reduction in sampling after both the button and the modal have been running live in production for a couple weeks. No rush on this. It's in sprint 69, probably toward the end of the sprint, but could conceivably move to sprint 70. We'll see.
Change 272724 abandoned by Bmansurov:
Reduce sampling rate for language switcher
Reason:
Resurrect if needed.