Update DevOps Score calculations
Problem to solve
DevOps score is currently calculating based on GitLab feature usage
across the product and how this data compares to other instances on usage ping:
DevOps Score displays the usage of GitLab’s major features on your instance over the last 30 days, averaged over the number of active users in that time period. It also provides a Lead score per feature, which is calculated based on GitLab’s analysis of top-performing instances based on usage ping data that GitLab has collected. Your score is compared to the lead score of each feature and then expressed as a percentage at the bottom of said feature. Your overall DevOps Score is an average of your feature scores. You can use this score to compare your DevOps status to other organizations.
Although GitLab feature usage is correlated with improved DevOps success, it is not actually a measure of it. DORA metrics (&4358 (closed)) are the industry standard for measuring DevOps success:
Proposal
- Figure out a new way to calculate DevOps Score using a combination of
DORA metrics
ANDfeature usage
- Score should be out of 100%
- Allow this score to be calculated at the
Group
,Project
&Segment
-levels - Use score to show trends, compare teams and identify opportunities for improvement
Example calculation:
+ Deployment frequency: Elite = 1.0
+ Lead time: High = 0.75
+ Time to restore: Medium = 0.5
+ Change failure rate: Medium = 0.5
+ Feature usage: 0.7
Dev: 0.8
Ops: 0.7
Sec: 0.6
---------------
Total Avg: 69%