Skip to content

docs: post the AGE-100 after-number (stamped=0, n=0 over 6.3 days) - #195

Open
dzianisv wants to merge 1 commit into
mainfrom
age-100-after-number
Open

docs: post the AGE-100 after-number (stamped=0, n=0 over 6.3 days)#195
dzianisv wants to merge 1 commit into
mainfrom
age-100-after-number

Conversation

@dzianisv

Copy link
Copy Markdown
Owner

Closes the measurement side of AGE-100.

Aggregated all 143 hourly Waitlist Mailto Reconcile runs (dzianisv/VibeBrowserProductPage) since #237 shipped the app-version split (merged 2026-08-14T09:56:54Z) through 2026-08-20T16:02Z:

scanned=34 synced=0 skipped=34 failed=0 unstamped=0 stamped=0

on every single run, identical to the pre-release baseline. scanned counts all matching Chatwoot conversations ever (not just new ones since last run), so a constant 34 across 6.3 days means zero new mailto-fallback conversations of any kind arrived, not just zero stamped ones — down from ~1.9/day (20 of 21 in 11 days) before AGE-87 shipped.

Documented the honest caveat in the doc: stamped=0 is a pass on the stated AGE-100 condition, but it's 0 samples, not 0-of-many. No evidence of a regression; also no positive confirmation the retry queue specifically is what's suppressing mailto (the ~436-device sideload cohort never auto-updates, so its background rate should be independent of the v0.4.13 rollout — its own drop to zero is unexplained but not this ticket's scope).

Raw evidence (run IDs, per-run scanned= lines, gap analysis) available on request; not committed to the repo to keep the doc readable.

Aggregated all 143 hourly Waitlist Mailto Reconcile runs since PR #237
(dzianisv/VibeBrowserProductPage) shipped the app-version split
(2026-08-14T09:56:54Z merge) through 2026-08-20T16:02Z: every run printed
scanned=34 synced=0 skipped=34 failed=0 unstamped=0 stamped=0, identical
to the pre-release baseline.

scanned counts all matching conversations ever, not just new ones, so a
constant 34 means zero new mailto-fallback conversations of any kind (not
just zero stamped ones) arrived in 6.3 days -- down from ~1.9/day before
AGE-87. stamped/unstamped are computed only over synced conversations, so
0/0 is a pass on an empty sample, not 0-out-of-many. Documented that
caveat explicitly so it isn't misread later.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant