Deployment Journal — Operational Intelligence Layer

Production deployment of the Operational Intelligence Layer — the largest single capability addition to the AI Execution Lab platform since the Failure Archive launch. This journal covers what was shipped, the one pre-deploy incident that was caught and fixed, and the post-deploy verification state.

1. Pre-Deploy State

What was being shipped:

This deployment packaged the Operational Intelligence Layer — a set of runtime systems that give the platform self-awareness about its own operational health and execution patterns.

Key files added or significantly modified:

Total content items at deploy: 424 pages (up from ~380 at the previous major deploy)

Deployment target: Vercel (Production, main branch)

2. Build Results

▲ Vercel CLI 38.0.0
→ Running build in /vercel/path0

Installing dependencies...
✓ node_modules installed (node direct, npm blocked per project config)

Running next build...
  ▲ Next.js 15.2.4

  Creating an optimized production build...
  ✓ Compiled successfully in 34.1s

  Linting and checking validity of types...
  ✓ TypeScript: 0 errors

  Collecting page data...
  ✓ lib/operational-memory.ts: indexed 424 content items
  ✓ lib/failure-memory.ts: indexed 12 failure reports (9 resolved, 3 monitoring)
  ✓ lib/confidence-scoring.ts: scored 424 items (avg confidence: 0.74)
  ✓ lib/execution-pathways.ts: mapped 38 execution pathways

  Generating static pages (424/424)
  ✓ Static pages generated successfully

Build Output:
  Route (app)                               Size     First Load JS
  ┌ ○ /                                     14.2 kB        142 kB
  ├ ○ /ops                                  22.8 kB        198 kB
  ├ ● /logs/[slug]                          8.4 kB         136 kB
  ├ ● /failures/[slug]                      9.1 kB         137 kB
  └ ... (424 total routes)

  First Load JS shared by all: 98.3 kB

✓ Build completed in 47.3s
✓ Deployment ready

Build time: 47.3 seconds total (compile + static generation)
TypeScript errors: 0
Build errors: 0
Pages generated: 424

The increase from the previous baseline of ~380 pages accounts for: new ops infrastructure pages, new lab content, new log entries, and the expanded /ops route set.

3. Post-Deploy Verification

Verified immediately after deployment promotion to production:

Execution pathways: /ops page loads with pathway map rendered. 38 pathways indexed, all rendering with correct source/destination content links. No broken references.

Confidence scores: Confidence badges visible on content pages that have evidence frontmatter or linked failure reports. The confidence-badge.tsx component renders correctly with the verified, approximate, and reconstructed quality tiers.

Debug context panel: The debug-context-panel.tsx component renders on the /ops page in collapsed state by default. Expand/collapse toggle working. Panel shows operational metadata: page count, failure index state, pathway coverage.

Failure memory index: /ops shows 12 indexed failures: 9 resolved, 3 in monitoring state. The failure-memory.ts index correctly picks up all files in content/failures/.

Operational memory: All 424 content items indexed. The operational-memory.ts system reads frontmatter at build time — no runtime filesystem access, fully static.

4. Incidents

Pre-Deploy: MDX Parse Error — debuggingComplexity Prop

Status: Caught and fixed before deploy. No production impact.

During content authoring for a new failure report, an MDX parse error was triggered:

Error: Unexpected token in JSX expression
  at: <FailureAnalysis debuggingComplexity="low" ...>

The issue: debuggingComplexity="low" was being parsed as a malformed JSX expression because the attribute value was followed immediately by another prop that the MDX parser was reading as a continuation of the string. The fix was adding explicit whitespace and verifying the prop ordering in the FailureAnalysis component call.

Time to diagnose and fix: ~8 minutes. The error message was clear enough to pinpoint the location. No other content files were affected.

5. What Shipped

A summary of every meaningful change in this deployment:

New runtime systems:

• lib/operational-memory.ts — build-time content index with frontmatter extraction
• lib/failure-memory.ts — failure archive indexing with resolution and recurrence tracking
• lib/confidence-scoring.ts — content confidence scoring based on evidence quality signals
• lib/execution-pathways.ts — pathway mapping from problem states to resolved outcomes

New components:

• components/operational/confidence-badge.tsx — inline badge, renders on content pages
• components/operational/debug-context-panel.tsx — collapsible ops debug panel

Upgraded pages:

• app/ops/page.tsx — full operational dashboard, now shows live counts, pathway map, failure index, confidence distribution

Content:

• 30+ new content items across content/logs/, content/labs/, content/docs/
• All new items picked up by the operational memory index automatically

Page count: 380-range → 424 pages