Test Pages/Crawl

Crawl test branch

This branch generates a procedurally-varying tree so a single Crawl-type project can be re-run multiple times to exercise new / removed / failed page-status detection. The tree is up to 6 levels deep — one beyond the Crawl project max-depth of 5 — so a full-depth crawl provably reaches every spine node.

Child links (3)

alpha (stable)
beta (stable)
transient-x1w (transient)

Link types

stable — deterministic URL, always 200. Forms the persistent spine of the tree.
transient — URL contains a fresh random ID per render. Appears in one crawl, disappears in the next → exercises new and removed page status.
flaky — stable URL, but the destination 404s on ~35% of requests. Best surfaced via the schedule-panel Run Now path, which records per-page HTTP status.