Hi guys,
Any word from analytics as to when the Media Viewer metrics dashboards will be updated? Is the EventsLogging migration almost done? Now is the time when we need this data. :)
http://multimedia-metrics.wmflabs.org/dashboards/mmv
Also, do we have enough data now to define acceptance performance criteria for Media Viewer and complete card #149?
https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/149
Other research questions which we would like to answer based on pilot results include:
* how fast do images load? I think we now have enough data to make a statement, both on a global and local basis, with breakdowns for warm and cold caches, as well as large images. Want to take a first stab at it, or want me to?
* does performance improve with more users? It would be great if we could confirm our hypothesis that images load faster when more users are clicking on them.
* any slowdowns during peak hours? Do we have any data that would show what happens during peak hours? Would this require us to create hourly dashboards in a few pilot sites?
* is the overall performance acceptable? This question can be answered once we define acceptable performance criteria. It could also be confirmed to some extent by survey responses, particularly if we see a decrease in complaints about the speed.
To that end, I will do another sweep at latest survey results and compile more results tomorrow on this page: https://www.mediawiki.org/wiki/Multimedia/Media_Viewer/Survey
Look forward to discussing these questions together in coming days, based on the data now on hand.
Thanks,
Fabrice
_______________________________
Fabrice Florin Product Manager Wikimedia Foundation
db1047 (where all the event logging data is being replicated to) is still lagging and we'll update the thread on the analytics list when it's done.
In the meantime, you can switch your queries from s1 to a new box that's caught up already. Here's a similar change I submitted for the mobile web team to consider:
https://gerrit.wikimedia.org/r/#/c/130833/1/mobile/config.yaml
On Thu, May 1, 2014 at 11:37 AM, Fabrice Florin fflorin@wikimedia.orgwrote:
Hi guys,
Any word from analytics as to when the Media Viewer metrics dashboards will be updated? Is the EventsLogging migration almost done? Now is the time when we need this data. :)
http://multimedia-metrics.wmflabs.org/dashboards/mmv
Also, do we have enough data now to define acceptance performance criteria for Media Viewer and complete card #149?
https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/149
Other research questions which we would like to answer based on pilot results include:
- how fast do images load?
I think we now have enough data to make a statement, both on a global and local basis, with breakdowns for warm and cold caches, as well as large images. Want to take a first stab at it, or want me to?
- does performance improve with more users?
It would be great if we could confirm our hypothesis that images load faster when more users are clicking on them.
- any slowdowns during peak hours?
Do we have any data that would show what happens during peak hours? Would this require us to create hourly dashboards in a few pilot sites?
- is the overall performance acceptable?
This question can be answered once we define acceptable performance criteria. It could also be confirmed to some extent by survey responses, particularly if we see a decrease in complaints about the speed.
To that end, I will do another sweep at latest survey results and compile more results tomorrow on this page: https://www.mediawiki.org/wiki/Multimedia/Media_Viewer/Survey
Look forward to discussing these questions together in coming days, based on the data now on hand.
Thanks,
Fabrice
Fabrice Florin Product Manager Wikimedia Foundation
Dear Dan,
Thanks so much for your prompt response, much appreciated!
Gilles and/or Mark, what do you recommend we do to update our metrics dashboards quickly, during this critical time?
I also look forward to your thoughts on the research questions below, now that we have lots of pilot data to base our answers on.
-f
On May 1, 2014, at 8:40 AM, Dan Andreescu dandreescu@wikimedia.org wrote:
db1047 (where all the event logging data is being replicated to) is still lagging and we'll update the thread on the analytics list when it's done.
In the meantime, you can switch your queries from s1 to a new box that's caught up already. Here's a similar change I submitted for the mobile web team to consider:
https://gerrit.wikimedia.org/r/#/c/130833/1/mobile/config.yaml
On Thu, May 1, 2014 at 11:37 AM, Fabrice Florin fflorin@wikimedia.org wrote: Hi guys,
Any word from analytics as to when the Media Viewer metrics dashboards will be updated? Is the EventsLogging migration almost done? Now is the time when we need this data. :)
http://multimedia-metrics.wmflabs.org/dashboards/mmv
Also, do we have enough data now to define acceptance performance criteria for Media Viewer and complete card #149?
https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/149
Other research questions which we would like to answer based on pilot results include:
- how fast do images load?
I think we now have enough data to make a statement, both on a global and local basis, with breakdowns for warm and cold caches, as well as large images. Want to take a first stab at it, or want me to?
- does performance improve with more users?
It would be great if we could confirm our hypothesis that images load faster when more users are clicking on them.
- any slowdowns during peak hours?
Do we have any data that would show what happens during peak hours? Would this require us to create hourly dashboards in a few pilot sites?
- is the overall performance acceptable?
This question can be answered once we define acceptable performance criteria. It could also be confirmed to some extent by survey responses, particularly if we see a decrease in complaints about the speed.
To that end, I will do another sweep at latest survey results and compile more results tomorrow on this page: https://www.mediawiki.org/wiki/Multimedia/Media_Viewer/Survey
Look forward to discussing these questions together in coming days, based on the data now on hand.
Thanks,
Fabrice
Fabrice Florin Product Manager Wikimedia Foundation
_______________________________
Fabrice Florin Product Manager Wikimedia Foundation
On Thu, May 01, 2014 at 11:40:08AM -0400, Dan Andreescu wrote:
https://gerrit.wikimedia.org/r/#/c/130833/1/mobile/config.yaml
I've done this on stat1003; we don't have any of yer high-falutin' config files in the multimedia analytics repos.
Thanks for the suggestion :)
Hi Mark,
Thanks so much for applying Dan’s proposed fix so quickly!
As discussed, I am puzzled by the numbers on these updated graphs:
http://multimedia-metrics.wmflabs.org/dashboards/mmv#actions-graphs-tab
Global image views used to be up above 3 million on Sunday 4/27, after the launch to 9 new pilot sites but before the outage (1). However, they now only appear to be 2 million. And the decline is surprising, compared to other pilot launches where image views keep increasing.
Also, local image views on the Polish Wikipedia used to be above 1 million on the same date before the outage, but are now down to 500k for the same date:
http://multimedia-metrics.wmflabs.org/dashboards/mmv_plwiki
Dan, Gilles or others, any ideas what’s going on — and how we can get the metrics back on track? I’m Cc:ing Analytics in case others have some insights that could help us.
We’re launching today on French and Dutch Wikipedias, and would like to have stable metrics to discuss with our communities.
Thanks for any help you can provide. :)
Fabrice
(1) https://www.mediawiki.org/wiki/Multimedia/Media_Viewer/Release_Plan#Pilots
On May 1, 2014, at 9:19 AM, Mark Holmquist mtraceur@member.fsf.org wrote:
On Thu, May 01, 2014 at 11:40:08AM -0400, Dan Andreescu wrote:
https://gerrit.wikimedia.org/r/#/c/130833/1/mobile/config.yaml
I've done this on stat1003; we don't have any of yer high-falutin' config files in the multimedia analytics repos.
Thanks for the suggestion :)
-- Mark Holmquist Software Engineer, Multimedia Wikimedia Foundation mtraceur@member.fsf.org https://wikimediafoundation.org/wiki/User:MHolmquist _______________________________________________ Multimedia mailing list Multimedia@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/multimedia
On May 1, 2014, at 8:40 AM, Dan Andreescu dandreescu@wikimedia.org wrote:
db1047 (where all the event logging data is being replicated to) is still lagging and we'll update the thread on the analytics list when it's done.
In the meantime, you can switch your queries from s1 to a new box that's caught up already. Here's a similar change I submitted for the mobile web team to consider:
https://gerrit.wikimedia.org/r/#/c/130833/1/mobile/config.yaml
On Thu, May 1, 2014 at 11:37 AM, Fabrice Florin fflorin@wikimedia.org wrote: Hi guys,
Any word from analytics as to when the Media Viewer metrics dashboards will be updated? Is the EventsLogging migration almost done? Now is the time when we need this data. :)
http://multimedia-metrics.wmflabs.org/dashboards/mmv
Also, do we have enough data now to define acceptance performance criteria for Media Viewer and complete card #149?
https://wikimedia.mingle.thoughtworks.com/projects/multimedia/cards/149
Other research questions which we would like to answer based on pilot results include:
- how fast do images load?
I think we now have enough data to make a statement, both on a global and local basis, with breakdowns for warm and cold caches, as well as large images. Want to take a first stab at it, or want me to?
- does performance improve with more users?
It would be great if we could confirm our hypothesis that images load faster when more users are clicking on them.
- any slowdowns during peak hours?
Do we have any data that would show what happens during peak hours? Would this require us to create hourly dashboards in a few pilot sites?
- is the overall performance acceptable?
This question can be answered once we define acceptable performance criteria. It could also be confirmed to some extent by survey responses, particularly if we see a decrease in complaints about the speed.
To that end, I will do another sweep at latest survey results and compile more results tomorrow on this page: https://www.mediawiki.org/wiki/Multimedia/Media_Viewer/Survey
Look forward to discussing these questions together in coming days, based on the data now on hand.
Thanks,
Fabrice
Fabrice Florin Product Manager Wikimedia Foundation
_______________________________
Fabrice Florin Product Manager Wikimedia Foundation
As discussed, I am puzzled by the numbers on these updated graphs:
I wish I had time to look into this more thoroughly but I'm swamped today. Also, most of the analytics team will be in Zurich starting Monday and we'll only be back after May 12th.
The best advice I can offer is to compare data from the two databases for the graphs where you see discrepancies. It's possible that replication lag / problems are still happening. Please report any bugs in bugzilla under the EventLogging component.
I have cc-ed the eventlogging-alerts list. I think the best path to go forward is to check back your numbers once an alert has been sent to the list saying that replication issues are resolved.
As Dan said most of analytics team should be traveling Monday but ops should be able to send a note to the list once replication issues are mostly resolved.
Thanks,
Nuria
On Thu, May 1, 2014 at 6:51 PM, Dan Andreescu dandreescu@wikimedia.orgwrote:
As discussed, I am puzzled by the numbers on these updated graphs:
I wish I had time to look into this more thoroughly but I'm swamped today. Also, most of the analytics team will be in Zurich starting Monday and we'll only be back after May 12th.
The best advice I can offer is to compare data from the two databases for the graphs where you see discrepancies. It's possible that replication lag / problems are still happening. Please report any bugs in bugzilla under the EventLogging component.
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
Thanks, Nuria,
Much appreciated.
Look forward to getting notified when replication issues are resolved.
We’re still puzzling over the strange dip shown on this graph, which we cannot explain, after checking with community members in the current pilot sites.
http://multimedia-metrics.wmflabs.org/dashboards/mmv
Last weekend, this graph showed 3 million image views globally, and was growing steadily. Any explanations of why it’s dropped down so much would be much appreciated, as the current data don’t appear to match what we’re hearing from our customers.
Have a safe trip!
Fabrice
On May 4, 2014, at 9:11 AM, Nuria Ruiz nuria@wikimedia.org wrote:
I have cc-ed the eventlogging-alerts list. I think the best path to go forward is to check back your numbers once an alert has been sent to the list saying that replication issues are resolved.
As Dan said most of analytics team should be traveling Monday but ops should be able to send a note to the list once replication issues are mostly resolved.
Thanks,
Nuria
On Thu, May 1, 2014 at 6:51 PM, Dan Andreescu dandreescu@wikimedia.org wrote: As discussed, I am puzzled by the numbers on these updated graphs:
I wish I had time to look into this more thoroughly but I'm swamped today. Also, most of the analytics team will be in Zurich starting Monday and we'll only be back after May 12th.
The best advice I can offer is to compare data from the two databases for the graphs where you see discrepancies. It's possible that replication lag / problems are still happening. Please report any bugs in bugzilla under the EventLogging component.
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
Multimedia mailing list Multimedia@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/multimedia
_______________________________
Fabrice Florin Product Manager Wikimedia Foundation
Fabrice,
Issues with event logging replication were resolved, the data should now accurately reflect your usage.
Thanks,
Nuria
On Sun, May 4, 2014 at 8:49 PM, Fabrice Florin fflorin@wikimedia.orgwrote:
Thanks, Nuria,
Much appreciated.
Look forward to getting notified when replication issues are resolved.
We’re still puzzling over the strange dip shown on this graph, which we cannot explain, after checking with community members in the current pilot sites.
http://multimedia-metrics.wmflabs.org/dashboards/mmv
Last weekend, this graph showed 3 million image views globally, and was growing steadily. Any explanations of why it’s dropped down so much would be much appreciated, as the current data don’t appear to match what we’re hearing from our customers.
Have a safe trip!
Fabrice
On May 4, 2014, at 9:11 AM, Nuria Ruiz nuria@wikimedia.org wrote:
I have cc-ed the eventlogging-alerts list. I think the best path to go forward is to check back your numbers once an alert has been sent to the list saying that replication issues are resolved.
As Dan said most of analytics team should be traveling Monday but ops should be able to send a note to the list once replication issues are mostly resolved.
Thanks,
Nuria
On Thu, May 1, 2014 at 6:51 PM, Dan Andreescu dandreescu@wikimedia.orgwrote:
As discussed, I am puzzled by the numbers on these updated graphs:
I wish I had time to look into this more thoroughly but I'm swamped today. Also, most of the analytics team will be in Zurich starting Monday and we'll only be back after May 12th.
The best advice I can offer is to compare data from the two databases for the graphs where you see discrepancies. It's possible that replication lag / problems are still happening. Please report any bugs in bugzilla under the EventLogging component.
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
Multimedia mailing list Multimedia@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/multimedia
Fabrice Florin Product Manager Wikimedia Foundation
http://en.wikipedia.org/wiki/User:Fabrice_Florin_(WMF)
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
Dear Nuria,
Thanks so much for this confirmation, much appreciated. :)
I would like to take this opportunity to thank everyone on the analytics team for your wonderful support of our multimedia work!
We just had a sprint retrospective today, and our team members were very grateful for your swift response to our questions, even while you were all meeting offsite in Switzerland.
Thanks to you guys, we now have a helpful set of metrics dashboards to inform our work on the Media Viewer:
https://www.mediawiki.org/wiki/Multimedia/Metrics#Client-side
… and we hope to soon have the same level of analytics for Upload Wizard, our next big project. (Please let us know if there are any existing metrics that we should be aware of on that project).
Cheers,
Fabrice
On May 14, 2014, at 9:30 AM, Nuria Ruiz nuria@wikimedia.org wrote:
Fabrice,
Issues with event logging replication were resolved, the data should now accurately reflect your usage.
Thanks,
Nuria
On Sun, May 4, 2014 at 8:49 PM, Fabrice Florin fflorin@wikimedia.org wrote: Thanks, Nuria,
Much appreciated.
Look forward to getting notified when replication issues are resolved.
We’re still puzzling over the strange dip shown on this graph, which we cannot explain, after checking with community members in the current pilot sites.
http://multimedia-metrics.wmflabs.org/dashboards/mmv
Last weekend, this graph showed 3 million image views globally, and was growing steadily. Any explanations of why it’s dropped down so much would be much appreciated, as the current data don’t appear to match what we’re hearing from our customers.
Have a safe trip!
Fabrice
On May 4, 2014, at 9:11 AM, Nuria Ruiz nuria@wikimedia.org wrote:
I have cc-ed the eventlogging-alerts list. I think the best path to go forward is to check back your numbers once an alert has been sent to the list saying that replication issues are resolved.
As Dan said most of analytics team should be traveling Monday but ops should be able to send a note to the list once replication issues are mostly resolved.
Thanks,
Nuria
On Thu, May 1, 2014 at 6:51 PM, Dan Andreescu dandreescu@wikimedia.org wrote: As discussed, I am puzzled by the numbers on these updated graphs:
I wish I had time to look into this more thoroughly but I'm swamped today. Also, most of the analytics team will be in Zurich starting Monday and we'll only be back after May 12th.
The best advice I can offer is to compare data from the two databases for the graphs where you see discrepancies. It's possible that replication lag / problems are still happening. Please report any bugs in bugzilla under the EventLogging component.
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
Multimedia mailing list Multimedia@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/multimedia
Fabrice Florin Product Manager Wikimedia Foundation
http://en.wikipedia.org/wiki/User:Fabrice_Florin_(WMF)
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
Multimedia mailing list Multimedia@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/multimedia
_______________________________
Fabrice Florin Product Manager Wikimedia Foundation
On Thu, May 1, 2014 at 8:37 AM, Fabrice Florin fflorin@wikimedia.orgwrote:
- how fast do images load?
I think we now have enough data to make a statement, both on a global and local basis, with breakdowns for warm and cold caches, as well as large images. Want to take a first stab at it, or want me to?
It seems that a lot of data is missing after April 27, but that probably won't affect averages much, so we can answer this one.
For image hits (i.e. when no thumbnail rendering is needed) an average image request takes 0.4 second. 10% of the users need to wait more than 2 sec, and 1% need to wait longer than 10 sec. For image misses, the average is 1.5 second, 10% of the users need to wait more than 5 sec. In the slowest 1% there is much larger variation than for hits, but they usually have to wait more than 20 sec.
We do not have a breakdown per image size, although we do have the data required to do it.
- does performance improve with more users?
It would be great if we could confirm our hypothesis that images load faster when more users are clicking on them.
The miss ratio has been climbing down steadily. We don't have a chart showing that (we probably should), but here is a Google Charts charthttps://chart.googleapis.com/chart?cht=lc&chs=600x400&chds=0,1&chxt=x,y&chxr=1,0,100|0,1,31,1&chxl=1:Apr+1|11|21|May+1&chd=t:0.7744107744,0.7146892655,0.6157407407,0.6258234519,0.6097087379,0.6268939394,0.6531007752,0.662027833,0.5877192982,0.6180371353,0.6314102564,0.6047297297,0.5735849057,0.618705036,0.5930232558,0.4857612267,0.3755687784,0.2952091255,0.2792185921,0.3090937403,0.2896433741,0.2661846309,0.2562147829,0.251244208,0.2177919249,0.232136633,0.2255857954,0.2392075695,0.2198016295,0.2327399767,0.2292069632 showing the ratio of misses within all requests in April - it started at 80%, and seems to have stabilized now a little above 20%.
Accordingly, the mean image load time is down from 1+ sec to 0.5 sec, the time for the 90th percentile is down from 4.5 sec to 3 sec, the 99th percentile is around 15 sec (there is no old number as there wasn't enough data before enabling on plwiki so the variation was huge).
* any slowdowns during peak hours?
Do we have any data that would show what happens during peak hours? Would this require us to create hourly dashboards in a few pilot sites?
We only have daily graphs currently. We do have the data, but we would have to write new queries.
multimedia@lists.wikimedia.org