Skip to content

[docs-agent] Add Try It method audit (DOCS-101)#1340

Closed
alchemy-bot wants to merge 1 commit into
mainfrom
docs/method-audit-catalog
Closed

[docs-agent] Add Try It method audit (DOCS-101)#1340
alchemy-bot wants to merge 1 commit into
mainfrom
docs/method-audit-catalog

Conversation

@alchemy-bot
Copy link
Copy Markdown
Contributor

Summary

Adds a categorized catalog of every failing "Try It" example across the JSON-RPC OpenRPC specs under src/openrpc/ (chain specs + Alchemy product specs).

This was requested in the #dx-docs-feedback Slack thread where Caleigh flagged that trace_* and debug_* Try It examples were failing for ~50% of methods. Rather than fix in isolation, this catalog probes every method's literal example payload against the documented endpoint and groups the failures so cluster fixes can be prioritized.

Headline numbers

  • Probed: 3,749 methods across 91 specs
  • OK: 2,604
  • Failed: 842 (24.4%)
  • Skipped (websocket-only or no example): 303

Three single-edit quick wins called out in the report

  1. eth_getProof shared example has an invalid hex character0x283s34c8e2b1... includes a literal s, breaking the example on ~29 EVM chains. One edit in src/openrpc/chains/_components/custom/methods.yaml.
  2. eth_submitWork is referenced from ~64 chain specs that aren't PoW — every one returns does not exist/is not available.
  3. eth_getAccount (Erigon-only) is referenced from ~44 chain specs that don't run Erigon — same shape.

The full categorized breakdown, per-spec failure rates, and methods-failing-on-N+-chains tables are in audits/2026-05-29-try-it-audit/report.md. failures.csv has one row per failure for spreadsheet pivot work. Scripts to reproduce are checked in alongside.

What this PR does NOT do

  • Fix any of the listed failures (separate cluster tickets after this catalog lands).
  • Audit REST OpenAPI specs (src/openapi/). Needs a per-spec adapter; flagged as follow-up.
  • Pull production logs of customer-hit errors (the other half of the original Slack ask). Needs data-team observability access.
  • Simulate Fern's Try It auto-fill on optional fields not in the example. The numbers in this catalog are therefore a lower bound on what users actually see.

Linear

DOCS-101 — Full Try It audit: catalog failing examples across Node API + REST specs

Requested by

@0xmts (via Slack thread)

Live-probes every method's canonical example payload across the 91 OpenRPC specs in src/openrpc/, classifies the failures, and surfaces three single-edit quick wins. Catalog lives at audits/2026-05-29-try-it-audit/.

Headline numbers: 3,749 probed, 2,604 ok, 842 failed (24.4%), 303 skipped (websocket-only or no example).

Refs DOCS-101

Requested-by: @0xmts
@alchemy-bot alchemy-bot requested a review from a team as a code owner May 29, 2026 17:27
@github-actions
Copy link
Copy Markdown

github-actions Bot commented May 29, 2026

🔗 Preview Mode

Name Status Preview Updated (UTC)
Alchemy Docs ✅ Ready 🔗 Visit Preview May 29, 2026, 5:28 PM

Copy link
Copy Markdown

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: c4d6cb10b0

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

Comment thread audits/2026-05-29-try-it-audit/scripts/run_audit.py
@SahilAujla
Copy link
Copy Markdown
Collaborator

good report but we don't wanna merge it, we can mark this task done!

@SahilAujla SahilAujla closed this May 29, 2026
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.

2 participants