Over-estimate Non-Viable SMAU Metrics
We're defining some stages as without viable metrics for SMAU purposes. Those are:
- Defend
- Monitor
- Configure
- Package
This issue outlines a proposal for over-estimating their SMAU numbers. The principle is that these stages all perform actions valuable to all users in a project. Below we'll highlight for each stage what those actions are and a proposal to capture their over-estimated SMAU with as little over-estimation as possible based on what we collect in Usage Ping today.
General Proposal
Known Passive Activation Tag (like number of projects with clusters added) * active_user_count (or user event counter) / count(Project)
This would effectively describe, what percentage of users are members of projects with the features activated.
Monitor Proposal
All active project users for any project with Prometheus enabled.
While we wait for other user-fied north start metrics
Configure
All active project user for any project with a cluster attached
While we wait for users creating pipeline deploys
Package
Count all users on instances with container registry enabled