Skip to content

Data Request: data regarding the effectiveness of the 2-engineer verification requirement

Description

The groupthreat insights and groupsecurity policies groups have an explicit and rigid verification process.

It was added with no data justifying it and also without a metric to track its usefulness or effectiveness.

groupcompliance has a similar "2 engineer verification" process, though it is more flexible as it allows for the use of the ~verified-by-author label

Similar processes is now being rolled out to new teams, without any way to track or measure the impact of such processes.

I would like to remove or loosen this process, or at least define a metric to track its usefulness.

I have an draft MR to iterate on that, collect data, etc.

As part of that effort I would like some help gathering analytics to compare:

Data requested

the data I would love to have is the historical data that covers prior to this process and after it:

  • issues completed/missed milestone
  • bug counts
  • average time to close issue
  • these numbers compared with dev teams that do not have this formal "a separate engineer must run validation"

the process was first added in September of 2020 (gitlab-com/www-gitlab-com@ee9e0e4c)

I am also open to any analytics that you think may be useful to determine if this process is a useful one

Acceptance Criteria / Functional Requirements

  • Technical requirements
  • fill in your requirement here
  • Interface requirements
  • fill in your requirement here
  • Quality requirements
  • fill in your requirement here
  • Performance requirements
  • fill in your requirement here

Implementation Details

High level task breakdown

  • fill in task breakdown here
Edited by Michael Becker