2.6 KiB
| description | stage | group | info |
|---|---|---|---|
| Learn how long your open merge requests have spent in code review, and what distinguishes the longest-running. | Manage | Optimize | To determine the technical writer assigned to the Stage/Group associated with this page, see https://about.gitlab.com/handbook/engineering/ux/technical-writing/#assignments |
Code Review Analytics (PREMIUM)
- Introduced in GitLab 12.7.
- Moved to GitLab Premium in 13.9.
Use Code Review Analytics to view the longest-running reviews among open merge requests, and:
- Take action on individual merge requests.
- Reduce overall cycle time.
NOTE: Initially, no data appears. Data is populated as users comment on open merge requests.
Overview
Code Review Analytics is available to users with Reporter access and above, and displays a table of open merge requests that have at least one non-author comment. The review time is measured from the time the first non-author comment was submitted.
To access Code Review Analytics, from your project's menu, go to Project Analytics > Code Review.
You can filter the list of merge requests by milestone and label.
The table is sorted by:
- Review time: Helping you to quickly find the longest-running reviews which may need intervention or to be broken down into smaller parts.
- Other columns: Display the author, approvers, comment count, and line change (-/+) counts.
Use cases
This feature is designed for development team leaders and others who want to understand broad code review dynamics, and identify patterns to explain them.
You can use Code Review Analytics to:
- Expose your team's unique challenges with code review.
- Identify improvements that might substantially accelerate your development cycle.
- Your team agrees that code review is moving too slow.
- The Value Stream Analytics feature shows that reviews are your team's most time-consuming step.
- Analyze the patterns and trends of different types of work that are moving slow.
For example:
- Lots of comments or commits? Maybe the code is too complex.
- A particular author is involved? Maybe more training is required.
- Few comments and approvers? Maybe your team is understaffed.