[
https://issues.apache.org/jira/browse/SOLR-15189?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17305989#comment-17305989
]
Jan Høydahl commented on SOLR-15189:
------------------------------------
Thanks for doing this Alexandre.
Guess this gives us some insight into traffic patterns. Would be interesting to
see the source of the refguide 6_6 traffic, but I'm quite convinced it is
Google, since 6.6 guide still shows up on top for some queries at my end.
We should prepare for sunsetting GA on both Lucene and Solr sites, note board
member Justin's comment on lucene dev list
https://lists.apache.org/thread.html/re44bf57334ca786b3b5c7c66f27c45604e15f9850920dd565ad64889%40%3Cdev.lucene.apache.org%3E
Also, I just noticed that INFRA must have started tracking downloads again,
since numbers have started appearing in
https://uls.apache.org/exports/solr.apache.org.yaml. Those numbers are so low
that I wonder if they are counting the right thing. My guess is that Solr
numbers are hidden in Lucene's stats since the current artifacts are actually
in lucene/solr folder of download site.
Of course any download numbers from this stat will only be the number of clicks
from homepage to the mirrors - and we don't know anything about how many of
those that actually ends up as a download, and we cannot know how many visit
the mirrors directly the next time. Also, with Docker rising the number of
Docker image pulls are euqally and increasingly interesting. We can get a count
of total pulls with this command
{{curl -s https://hub.docker.com/v2/repositories/library/solr/ | jq -r
".pull_count"}}
I have created SOLR-15275 for getting rid of GA again. Can we perhaps set April
1st as a date for that? Then we have about 1 month stats in GA to continue
analyzing.
> Tracking downloads on new Solr site
> -----------------------------------
>
> Key: SOLR-15189
> URL: https://issues.apache.org/jira/browse/SOLR-15189
> Project: Solr
> Issue Type: Sub-task
> Reporter: Jan Høydahl
> Assignee: Alexandre Rafalovitch
> Priority: Major
> Attachments: Analytics All Web Site Data All Traffic
> 20210219-20210320.pdf, Analytics All Web Site Data Pages 20210219-20210320.pdf
>
>
> On lucene.apache.org we use Google Analytics tracking
> {quote}GOOGLE_ANALYTICS_TRACKING_ID = 'UA-94576-12'
> {quote}
> I think the reason was so that we could estimate downloads from mirrors, by
> counting number of clicks on the links from download pages. But are anyone
> ever looking at or publishing those numbers?
> The ASF wants projects to stop using 3rd party tracking of users and instead
> ask INFRA for aggregated stats for the page. WDYT? Should we
> # Remove trackers from both sites and rely on stats from infra
> # Continue using Google analytics, but have someone actually publish numbers
> from it every month?
> # Use some other way of counting downloads?
> h2. What do we get without a tracker?
> INFRA provides anonymous page view stats here
> [https://uls.apache.org/exports/lucene.apache.org.yaml] which gives some
> insight. But not downloads specifically. We see 12k visits to Solr downloads
> page last months, but we don't know how many of those clicked...
> {code:java}
> Sheet3:
> Name: Most visited pages, past month
> Values:
> /solr/index.html: 33604
> /index.html: 27588
> /solr/downloads.html: 12118
> /core/2_9_4/queryparsersyntax.html: 11135
> /core/index.html: 10353
> /solr/guide/solr-tutorial.html: 9734
> /solr/resources.html: 8014
> /solr/features.html: 7046
> /solr/guide/8_8/solr-tutorial.html: 6099
> /solr/news.html: 5843
> /solr/guide/6_6/the-standard-query-parser.html: 5216
> /solr/guide/index.html: 4430
> /solr/guide/6_6/common-query-parameters.html: 4379
> /core/downloads.html: 3644
> {code}
> There's an interesting section at the bottom of that YAML page, wonder if it
> could be enabled in some way
> {code}
> Sheet6:
> Name: Downloads, past month
> Values: {}
> {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
