1514542 - [Growth] Stubinstaller and Installer Ping Documentation

Status: RESOLVED DUPLICATE of bug 1522363

(Data Science :: General, task)

Description

[Su-Young Hong]

Brief description of the request:

* DTMO documentation of the install data sources
* need to investigate and find out how they work
    - the fields
    - the sources (where they come from) 
    - differences in specific datasets that look the same

ALSO: 
* get OS added to firefox_installer_install_parquet, and possibly others
* actually, find any deficiencies with the data, file bugs for them (the above being an obvious one) 

Link to any assets:

*	firefox_installer_install_parquet (Athena/Presto)
	download_stats_year (DSMO-RS)
            download_stats_funnelcake
            download_stats

Is there a specific data scientist you would like or someone who has helped to triage this request:

* filed by self, assigned to self

Comment 1

[Su-Young Hong]

reference:

download stats (no longer updated)
download stats year (updated more frequently)
download_stats_funnelcake
- ping_version: contains info on funnelcake version (ex: v7-26)

Comment 2

[Su-Young Hong]

reference:

the SQL for DSMO-RS DB is redshift / postgres

Comment 3

[Su-Young Hong]

Bug Summary:

Research and write documentation for Stubinstaller and Installer pings, including but not limited to:

• coverage for FF versions, stubinstaller/installer versions, time periods served
  • broken down by OS types and OS architecture types (32 vs 64 bit)
• ping schema
• ping send conditions
• stubinstalling logic
• database management
  • in terms of how the pings get rolled up into derived tables, what they cover, etc.

Adjacent documentation for understanding what our web properties (mozilla.org and FTP sites) were serving to whom and when

• i.e. for Win users, we served a stubinstaller in X, Y, Z conditions starting on YYYY-MM-DD.

Adjacent documentation for how funnelcake stubinstaller pings and how to interpret

Next steps:

• get answers to above questions (find who we need to talk to as the source of truth for a lot of these)
• organize into good documentation
• put it into DTMO (I think this is where this should live)

Comment 4

[Su-Young Hong]

Note: so much of this work (documenting stubinstaller pings) has been done in Bug 1522363 by Emily.

The current state of that documentation is here:

• https://docs.telemetry.mozilla.org/datasets/other/stub_installer/reference.html

Going to keep this bug open but shift to scope to fill in additional details into this documentation.

Comment 5

[Su-Young Hong]

Current documentation is sufficient. If there is need for:

• contextualize stubinstaller events in Firefox onboarding lifecycle
• example queries
• comparison of downloads, installs, new profiles

In the future, we can reopen this / file a new bug