Daily Bluesky bot for AT Mot. Invites players and congratulates yesterday's solvers.
docs: correct SOLVER_SAMPLE_CAP subrequest-cost guidance
The README told readers to raise SOLVER_SAMPLE_CAP to ~200 on the paid
plan, directly contradicting the 50-subrequest budget it had just
described: at ~2 subrequests per sampled record, 200 records is ~400
subrequests. Replace the figure with the per-record cost so the cap is
sized against the actual budget, and mirror the note in config.ts.
author
Jeremy Herve
date
(Jun 25, 2026, 4:08 PM +0200)
commit
b88c8ce7
b88c8ce755aee1b519fa24e643b855f8dc90a569
parent
e94d6236
e94d62366e2669351fab0149597b955dec1e6c19
+14
-2
+8
-1
README.md
+8
-1
README.md
···
52
52
The free Workers plan is sufficient: each language runs in its own scheduled
53
53
invocation, so each gets the full 50-subrequest budget. Solver counting samples up
54
54
to `SOLVER_SAMPLE_CAP` (20) records per language; beyond that the count is hedged
55
-
(e.g. "20+"). On the paid plan, raise `SOLVER_SAMPLE_CAP` in `src/config.ts` to ~200.
55
+
(e.g. "20+").
56
+
57
+
Each sampled record costs ~2 subrequests (DID-document resolution + the record
58
+
read), so the default cap of 20 already uses most of the per-invocation budget on
59
+
the free plan — keep headroom for the count, session, idempotency check, and
60
+
publish (~6 more). On the paid plan you can raise `SOLVER_SAMPLE_CAP` in
61
+
`src/config.ts`, but size it against the subrequest cost (~2 per record), not just
62
+
the desired sample size.
+6
-1
src/config.ts
+6
-1
src/config.ts
···
38
38
export const APPVIEW_URL = 'https://public.api.bsky.app';
39
39
export const USER_AGENT = `atmot-bot/1.0 (+${ORIGIN}; daily word game bot)`;
40
40
41
-
/** Free Workers plan: keep ~2 subrequests/record under the 50/invocation cap. */
41
+
/**
42
+
* Solver sample size. Each record costs ~2 subrequests (DID-doc resolution +
43
+
* record read), so 20 records ≈ 40, leaving ~10 of the free plan's 50/invocation
44
+
* for the count, session, idempotency check, and publish. Raise with care, and
45
+
* size it against the ~2-subrequests-per-record cost — not just the sample size.
46
+
*/
42
47
export const SOLVER_SAMPLE_CAP = 20;
43
48
44
49
/** The result record shape we read from each player's PDS (colours only; subset we use). */