[jbosstools-issues] [JBoss JIRA] (JBIDE-23007) Track rh-eclipse46-devstudio-*.rpm version usage

[Nick Boldt (JIRA)]

    [ https://issues.jboss.org/browse/JBIDE-23007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13279355#comment-13279355 ] 

Nick Boldt edited comment on JBIDE-23007 at 8/15/16 7:09 PM:
-------------------------------------------------------------

If we rebuild (or simply tweak) the update site metadata we could embed a different "phone home" URL, via p2StatsURL:

https://github.com/jbdevstudio/jbdevstudio-product/blob/master/pom.xml#L118

Then we could look at those 404 requests to https://devstudio.redhat.com/usage/installs/$\{JOB_NAME}/$\{project.version}/$\{BUILD_ALIAS}/$\{buildQualifier}/ in the apache log to see how many installs come from rpm vs. from update site.

I don't believe we currently differentiate between the two options in #3 above (a., jar installer vs. b., p2/BYOE install vs., c., Marketplace install), nor would we be able to differentiate #2 ("installed via p2") vs. #3b or #3c.

In other words, any p2 installation that points at the same update site, eg., https://devstudio.redhat.com/10.0/stable/updates/ , will appear to be the same install regardless of HOW the user did the install.

So, I propose these 3 cases instead:

1) Devstudio installed via rh-eclipse46-devstudio-*.rpm - track using p2StatsURL or by counting the number of times the rpm is downloaded
2) Devstudio installed via p2 (BYOE, Marketplace, or on top of rh-eclipse46-base-*.rpm) - track using p2StatsURL
3) Devstudio installed via installer jar - track by counting the number of times the jar is downloaded




was (Author: nickboldt):
If we rebuild (or simply tweak) the update site metadata we could embed a different "phone home" URL, via p2StatsURL:

https://github.com/jbdevstudio/jbdevstudio-product/blob/master/pom.xml#L118

Then we could look at those 404 requests to https://devstudio.redhat.com/usage/installs/${JOB_NAME}/${project.version}/${BUILD_ALIAS}/${buildQualifier}/ in the apache log to see how many installs come from rpm vs. from update site.

I don't believe we currently differentiate between the two options in #3 above (a., jar installer vs. b., p2/BYOE install vs., c., Marketplace install), nor would we be able to differentiate #2 ("installed via p2") vs. #3b or #3c.

In other words, any p2 installation that points at the same update site, eg., https://devstudio.redhat.com/10.0/stable/updates/ , will appear to be the same install regardless of HOW the user did the install.

So, I propose these 3 cases instead:

1) Devstudio installed via rh-eclipse46-devstudio-*.rpm - track using p2StatsURL or by counting the number of times the rpm is downloaded
2) Devstudio installed via p2 (BYOE, Marketplace, or on top of rh-eclipse46-base-*.rpm) - track using p2StatsURL
3) Devstudio installed via installer jar - track by counting the number of times the jar is downloaded



> Track rh-eclipse46-devstudio-*.rpm version usage
> ------------------------------------------------
>
>                 Key: JBIDE-23007
>                 URL: https://issues.jboss.org/browse/JBIDE-23007
>             Project: Tools (JBoss Tools)
>          Issue Type: Feature Request
>          Components: usage
>            Reporter: Alexey Kazakov
>            Assignee: Alexey Kazakov
>             Fix For: 4.4.1.Final
>
>
> We need to distinguish the following use cases:
> 1) Devstudio installed via RPM
> 2) Devstudio installed via P2 but on top of the "base" eclipse rpm
> 3) Devstudio installed traditionally (jar installer or p2)
> As a bonus we could want to track if the DTS + devstudio rpms were both installed.



--
This message was sent by Atlassian JIRA
(v6.4.11#64026)