Maintenance card, staff "All companies": painted from bounded chunks, equal to the whole answer
p11-page-maintenance C4 part 2. With the whole-answer request blocked, the card is painted from the four 25-building chunks alone, and it shows the same numbers as the whole answer.
29 Sept 2026 · branch perf/maintenance-card-chunk-page · captured against propflow-stage · at a3528fe7d3+local
Data note. Every name and balance below comes from propflow-stage, which is anonymized by construction (ADR-0097/0110 — phones in the impossible +1000 NPA, emails @example.test). The identical capture against prod would carry real tenants' names and balances and is not safe to attach anywhere, which is why the generator refuses to run against one.
1 · From the chunks alone
Staff on "All companies". This context blocks the page's whole-answer request (the one that carries period=), so every number on the Maintenance card comes from the chunk reads, only=maintenance&chunk=0..3. The dev server log shows all four chunks answered 200, and no whole-answer request reached the server from this context.
The count read here is measured, and the next section must show the same count.
What this does not show: the transient "So far: N of M buildings counted. Loading the rest…" line. On stage the four chunks finish before the capture tool's page-load wait returns, so the line is already gone when the shot is taken. Its text and placement are pinned by use-maintenance-card-chunks.test.tsx, which renders the real MaintenanceSection mid-count.
Asserted in the captured DOM — open work orders · CREATED (WEEK) · SELF-SERVE · TOP CATEGORIES. The running total and its progress line are on screen together.
2 · The whole answer, for comparison
2 · The whole answer, for comparison
A fresh context with nothing blocked: the provider's whole answer lands, the chunk poll switches off (the review fix in a3528fe7d3), and the card shows the same open count as the chunks-only section.
Asserted in the captured DOM — 921 · open work orders · CREATED (WEEK) · SELF-SERVE
