01 · the problem
Seeing the problem.
Engineering managers need quick answers on deployments, incidents and repositories, but that data usually lives in three separate tools.
The story · 4 chapters
02 · the build
What I built.
- Organisation overview with KPIs, trends and a searchable team table
- Side-by-side comparison of two or three teams
- Team and repository drill-downs, with date ranges kept in the URL
65 to 1
database round-trips to load the team table
4
levels of detail, from organisation to repository
300 ms
debounced search, one request per typing burst
2 to 3
teams compared side by side on shared axes
03 · under the hood
Making it work.
Database-side aggregation. One query of joined subqueries replaced about 65 per-team queries.
Server-driven table. Sorting and pagination run on the server, so the browser only gets the current page.
Reliable by design. Validated routes, typed API helpers, and real loading, empty and error states.
04 · the result
Succeeding.
Four connected levels of detail in one responsive interface, with every filter consistent across API, queries and UI.