Build success rate
94.2%
Rolling 7-day window. A build counts as successful when every invariant passes + Section P scenarios pass + the manifest emits.
Quality
How Burooj is performing
Weekly snapshot of Burooj's own engineering quality numbers, regenerated every Monday 00:00 UTC. Two surfaces, one promise: the changelog tells you what we shipped, this page tells you how well it's running.
Live status is published separately on the status page — industry pattern, not Burooj convention.
This week (2026-05-11)
P50 build time
7.6 min
Median wall-clock minutes from "Start build" to a verified deployed staging URL.
Section P pass rate
100.0%
15 end-to-end smokes that gate revenue. A regression here means a paying customer would see a broken build.
Invariant-checker veto rate
5.8%
Percent of builds the invariant checker stopped pre-deploy. Higher is better for safety, worse for throughput; the page shows both so you can judge.
Escalation-fix success
88.1%
When the first-pass code fails, the fix-loop escalates to a stronger model. This is the recovery rate.
Signed-manifest emission
100.0%
Percent of successful builds that emit a KMS-signed Verifiable Build Manifest. Sub-100% is acceptable during the canary; the target is 100%.
4-week history
| Week | Build success rate | P50 build time | Section P pass rate | Invariant-checker veto rate | Escalation-fix success | Signed-manifest emission |
|---|---|---|---|---|---|---|
| 2026-05-11 | 94.2% | 7.6 min | 100.0% | 5.8% | 88.1% | 100.0% |
| 2026-05-04 | 93.0% | 7.9 min | 100.0% | 6.4% | 87.5% | 100.0% |
| 2026-04-27 | 92.1% | 8.2 min | 100.0% | 6.9% | 86.3% | 100.0% |
| 2026-04-20 | 91.4% | 8.7 min | 100.0% | 7.2% | 84.6% | 100.0% |
Numbers come from the same internal data our team reads daily. The
weekly regeneration job lives at
.github/workflows/regenerate-quality-page.yml. If a number
looks wrong, the source is the same one we hold ourselves to — file
an issue.