Skip to content

docs(observe): split Observe + traceAI into product and SDK sections#681

Open
abhijaisrivastava15 wants to merge 3 commits into
devfrom
docs/observe-traceai-revamp
Open

docs(observe): split Observe + traceAI into product and SDK sections#681
abhijaisrivastava15 wants to merge 3 commits into
devfrom
docs/observe-traceai-revamp

Conversation

@abhijaisrivastava15

@abhijaisrivastava15 abhijaisrivastava15 commented Jun 23, 2026

Copy link
Copy Markdown

Summary

Restructures the Observe documentation into two top-level sections — Observe (the product) and traceAI (the instrumentation SDK) — and rewrites the pages to the product-docs-playbook bar. Graded page-by-page against a locked benchmark, the section moves from the 1.8/5 audit baseline → 4.40/5, with every page now at ≥4/5.

This is a fresh PR for the revamp (separate from #640).

Information architecture

  • Two sections: Observe (Overview → Getting Started → Concepts → Features → Reference → Troubleshooting) and a new top-level traceAI (Overview → Quick start → Concepts → Manual instrumentation → Framework integrations → Reference → Troubleshooting).
  • Manual instrumentation (14 pages) and Framework integrations moved under traceAI.
  • Redirects added for every moved slug — audit-links reports 0 broken links.

New pages

  • Observe: Getting Started, Quick start, Observability model, Tags.
  • traceAI: Overview, Quick start.

Rewrites to depth

  • Concepts (Traces, Spans, Sessions and users, OpenTelemetry, traceAI SDK) to the 9-section concept shape with Mermaid diagrams.
  • Feature deep-dives (Trace explorer, Dashboards, Tags, Sessions, Users, Alerts, Trace evals, Voice) with parameter tables, cropped synthetic-data screenshots, and troubleshooting.
  • Reference (Filter syntax, Dashboard metrics, Export and endpoints) completed with full contracts.
  • 41 framework-integration leaves rewritten to install → configure → run → verify → troubleshoot, each with what-spans-are-produced, a troubleshooting table, a Span-concept link, and full frontmatter.

Quality — graded against the benchmark

Pass Overall Observe traceAI
v1 3.60 3.79 3.18
v2 4.29 4.38 4.09
v3 4.40 4.50 4.18
v4 (final) 4.40 4.50 4.18

Final pass: every page ≥4/5, 0 pages 404, 0 broken links, build clean (1002 pages).

Reports

  • 📊 Observe-Docs-Analysis.pdf — the competitor/source analysis the quality bar is drawn from.
  • 📐 Benchmark.pdf — the locked answer key: page-type cards, playbook minimums, and the coverage map.
  • Scorecard-v4.pdf — every page graded page-by-page against the benchmark (final pass).

Click any report to open it in GitHub's PDF viewer.

Test plan

  • npm run build clean (1002 pages)
  • npm run audit-links — 0 broken nav, 0 broken content links
  • No banned words / TODO markers; every page passes its page-type frontmatter
  • Reviewer spot-check of rendered Observe + traceAI pages

Restructure the Observe docs into two top-level sections — Observe (the product)
and traceAI (the instrumentation SDK) — per the locked IA decision. Rewrite pages
to the product-docs-playbook bar, add new Getting Started, Quick start, concept,
and reference pages, move Manual instrumentation and Framework integrations under
traceAI, and add redirects so no existing URL 404s.
@entelligence-ai-pr-reviews

Copy link
Copy Markdown

Automatic Review Skipped

Too many files for automatic review.

If you would still like a review, you can trigger one manually by commenting:

@entelligence review

…ages

Apply review feedback on the Observe + traceAI sections:
- Remove the standalone Overview; make Get started the first sidebar item in
  both Observe and traceAI, for a consistent spine across sections.
- Open each landing with a plain "What is X?" that links straight to the
  trace and span concepts, and link the landing to the reference.
- Cut the Observe landing from ~880 to ~280 words to reduce overload; drop the
  unclear "Why this matters in production" heading and simplify the language.
…eview artifacts

Reference reports for the Observe + traceAI revamp: the source/competitor
analysis, the locked benchmark (answer key), and the final page-by-page
scorecard. Kept under pr-assets/ (outside src/ and public/, so not published
to the docs site).
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