Heatmap · Diagram Design

The week payments went dark

CI failure rate (%) by service and sprint, six sprints. payments spiked to 47% in S4 — a deploy regression that slipped through the gate — and had recovered to 4% by S6. billing carries the highest structural rate; notifications is the quietest service throughout.

The week payments went dark Heatmap of CI failure rate by service and sprint; payments service spiked to 47% in sprint 4 due to a deploy regression, recovered to 4% by sprint 6; billing carries the highest baseline failure rate at 9% in sprint 3. SERVICE SPRINT S1 S2 S3 S4 S5 S6 auth payments api-gateway billing notifications 4% 5% 6% 3% 2% 2% 3% 4% 5% 47% REGRESSION 6% 4% 2% 3% 4% 2% 1% 1% 5% 7% 9% 6% 4% 3% 1% 2% 2% 2% 1% 1% DEPLOY REGRESSION HIGH BASELINE LEGEND ILLUSTRATIVE CI FAILURE RATES · NOT A MEASURED REGISTRY LOW HIGH payments · S4 · 47% — deploy regression (now fixed) NON-FOCAL SCALE: 1–9% · SINGLE INK RAMP · ACCENT IS EDITORIAL, NOT A DATA RANGE

Focal · payments · S4

Deploy regression — 47%

A CI gate misconfiguration during the payments S4 deploy let a breaking change through. The spike resolved by S6 after the rollback and a gate fix, returning to a 4% baseline.

Structural issue · billing

Highest baseline — peaks 9%

billing shows the highest non-focal failure rates across all six sprints, peaking at 9% in S3. This is a persistent structural issue unrelated to the payments regression.

Quietest service · notifications

Most stable — never above 2%

notifications stayed at or below 2% every sprint. This is the contrast the heatmap makes visible: same period, same team, dramatically different structural rates.

Heatmap · Diagram Design Illustrative CI failure rates · not a measured registry