Commit Graph

4 Commits

Author SHA1 Message Date
Felipe Cardoso
88cf4e0abc feat: Update to production model stack and fix remaining inconsistencies
## Model Stack Updates (User's Actual Models)

Updated all documentation to reflect production models:
- Claude Opus 4.5 (primary reasoning)
- GPT 5.1 Codex max (code generation specialist)
- Gemini 3 Pro/Flash (multimodal, fast inference)
- Qwen3-235B (cost-effective, self-hostable)
- DeepSeek V3.2 (self-hosted, open weights)

### Files Updated:
- ADR-004: Full model groups, failover chains, cost tables
- ADR-007: Code example with correct model identifiers
- ADR-012: Cost tracking with new model prices
- ARCHITECTURE.md: Model groups, failover diagram
- IMPLEMENTATION_ROADMAP.md: External services list

## Architecture Diagram Updates

- Added LangGraph Runtime to orchestration layer
- Added technology labels (Type-Instance, transitions)

## Self-Hostability Table Expanded

Added entries for:
- LangGraph (MIT)
- transitions (MIT)
- DeepSeek V3.2 (MIT)
- Qwen3-235B (Apache 2.0)

## Metric Alignments

- Response time: Split into API (<200ms) and Agent (<10s/<60s)
- Cost per project: Adjusted to $100/sprint for Opus 4.5 pricing
- Added concurrent projects (10+) and agents (50+) metrics

## Infrastructure Updates

- Celery workers: 4-8 instances (was 2-4) across 4 queues
- MCP servers: Clarified Phase 2 + Phase 5 deployment
- Sync interval: Clarified 60s fallback + 15min reconciliation

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-29 23:35:51 +01:00
Felipe Cardoso
f138417486 fix: Resolve ADR/Requirements inconsistencies from comprehensive review
## ADR Compliance Section Fixes

- ADR-007: Fixed invalid NFR-501 and TC-002 references
  - NFR-501 → NFR-402 (Fault tolerance)
  - TC-002 → Core Principle (self-hostability)

- ADR-008: Fixed invalid NFR-501 reference
  - Added TC-006 (pgvector extension)

- ADR-011: Fixed invalid FR-201-205 and NFR-201 references
  - Now correctly references FR-401-404 (Issue Tracking series)

- ADR-012: Fixed invalid FR-401, FR-402, NFR-302 references
  - Now references new FR-800 series (Cost & Budget Management)

- ADR-014: Fixed invalid FR-601-605 and FR-102 references
  - Now correctly references FR-203 (Autonomy Level Configuration)

## ADR-007 Model Identifier Fix

- Changed "claude-sonnet-4-20250514" to "claude-3-5-sonnet-latest"
- Matches documented primary model (Claude 3.5 Sonnet)

## New Requirements Added

- FR-801: Real-time cost tracking
- FR-802: Budget configuration (soft/hard limits)
- FR-803: Budget alerts
- FR-804: Cost analytics

This resolves all HIGH priority inconsistencies identified by the
4-agent parallel review of ADRs against requirements and architecture.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-29 14:13:26 +01:00
Felipe Cardoso
de47d9ee43 fix: Resolve ADR-007 vs ADR-010 Temporal contradiction
Remove Temporal from the architecture in favor of the simpler
transitions + PostgreSQL + Celery approach. This aligns ADR-007
with ADR-010 based on user preference for simpler operations.

Key changes:
- ADR-007 now recommends transitions library instead of Temporal
- Added explicit "Why Not Temporal?" section explaining the trade-off
- Added "Reboot Survival" section documenting durability guarantees
- Updated architecture diagrams and component responsibilities
- Updated ARCHITECTURE.md summary matrix

The simpler approach is more appropriate for Syndarix's scale (10-50
concurrent agents) and uses existing PostgreSQL + Celery infrastructure.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-29 14:04:37 +01:00
Felipe Cardoso
bd702734c2 docs: add ADR-007 for agentic framework selection
Establishes the hybrid architecture decision:
- LangGraph for agent state machines (MIT, self-hostable)
- Temporal for durable workflow execution (MIT, self-hostable)
- Redis Streams for agent communication (BSD-3, self-hostable)
- LiteLLM for unified LLM access (MIT, self-hostable)

Key decision: Use production-tested open-source components rather than
reinventing the wheel, while maintaining 100% self-hostability with
no mandatory subscriptions.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2025-12-29 13:42:33 +01:00