Overview
How CI is doing across all repositories. Click a repository anywhere to focus the whole page on it.
Which repositories are fast and reliable?
One bubble per repository. Higher passes more often, further left finishes sooner, bigger runs more.
When do runs start?
Runs per hour of the day. The shaded band is 09:00–17:00.
How many runs pass each day?
Delivery (DORA)
The four standard measures of how well a team ships software, rated from Low to Elite.
How does each repository do?
When did each repository ship?
Each mark is a release or a deploy.
Inside a job
Pick a job to see where its time went and, if it failed, why.
Job
The job's log
The important lines: commands, errors, warnings and test results.
Names, emails and tokens are removed.
Health check
Why jobs fail, where their time goes, and whether the basic safety settings are in place.
Why do jobs fail?
Failed jobs from the log sample, grouped by cause.
Where does job time go?
Share of time in successful jobs, read from their logs.
Are the safety basics in place?
✓ fine ! worth a look ✕ fix this. Hover a mark for details.
Test audit
For one recent pull request per repository, every test from every job was matched, to find tests that ran more than once.
How often does each test run?
Test runs on one pull request commit. Grey is each test once; the rest are repeats.
What should change?
See every job of one audited commit
AI impact
How much AI writes, reviews and runs code. AI is only counted where it leaves a trace, so real use is higher.
Which AI tools write code?
Pull requests that name each tool.
Which AI models are named?
Pull requests that name each model in a commit.
Do AI-written changes cause more problems?
Share of pull requests whose first CI run failed, got "changes requested", or were reverted or fixed later.
Who starts CI runs?
Share of CI time.
How ready is each repository for AI?
Solution box
Every problem found, with a fix. Work through them in this order: