dullfig/xml-pipeline
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
143 commits 1 branch 0 tags 1.8 MiB
Commit graph

3 commits

Author SHA1 Message Date
dullfig
e6697f0ea2 Add BudgetWarning system alerts for token budget thresholds 
- Create BudgetWarning primitive payload (75%, 90%, 95% thresholds)
- Add threshold tracking to ThreadBudget with triggered_thresholds set
- Change consume() to return (budget, crossed_thresholds) tuple
- Wire warning injection in LLM router when thresholds crossed
- Add 15 new tests for threshold detection and warning injection

Agents now receive BudgetWarning messages when approaching their token limit,
allowing them to design contingencies (summarize, escalate, save state).

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2026-01-27 21:41:34 -08:00
dullfig
860395cd58 Add usage/gas tracking REST API endpoints 
Endpoints:
- GET /api/v1/usage - Overview with totals, per-agent, per-model breakdown
- GET /api/v1/usage/threads - List all thread budgets sorted by usage
- GET /api/v1/usage/threads/{id} - Single thread budget details
- GET /api/v1/usage/agents/{id} - Usage totals for specific agent
- GET /api/v1/usage/models/{model} - Usage totals for specific model
- POST /api/v1/usage/reset - Reset all usage tracking

Models:
- UsageTotals, UsageOverview, UsageResponse
- ThreadBudgetInfo, ThreadBudgetListResponse
- AgentUsageInfo, ModelUsageInfo

Also adds has_budget() method to ThreadBudgetRegistry for checking
if a thread exists without auto-creating it.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2026-01-27 21:20:36 -08:00
dullfig
8b11323a8b Add token budget enforcement and usage tracking 
Token Budget System:
- ThreadBudgetRegistry tracks per-thread token usage with configurable limits
- BudgetExhaustedError raised when thread exceeds max_tokens_per_thread
- Integrates with LLMRouter to block LLM calls when budget exhausted
- Automatic cleanup when threads are pruned

Usage Tracking (for production billing):
- UsageTracker emits events after each LLM completion
- Subscribers receive UsageEvent with tokens, latency, estimated cost
- Cost estimation for common models (Grok, Claude, GPT, etc.)
- Aggregate stats by agent, model, and totals

Configuration:
- max_tokens_per_thread in organism.yaml (default 100k)
- LLMRouter.complete() accepts thread_id and metadata parameters

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2026-01-27 21:07:43 -08:00