costs

Cost Report

Generate a cost report using OpenTelemetry metrics from the platform operations API.

Instructions

1. Fetch cost data from the ops API:

   • Use Bash to call the costs endpoint:

   curl -s http://backend:8000/api/ops/costs \
     -H "Authorization: Bearer $TRINITY_MCP_API_KEY" \
     -H "Content-Type: application/json"
   

2. Check the response:

   • If enabled: false - OTel is not configured, show setup instructions
   • If available: false - OTel is enabled but collector is unreachable
   • If available: true - Parse and display the metrics

3. Generate the report based on the response:

When OTel is Available

## Platform Cost Report
Generated: {timestamp}

### Summary
- Total Cost: ${summary.total_cost} USD
- Total Tokens: {summary.total_tokens:,}
- Daily Limit: ${summary.daily_limit} ({summary.cost_percent_of_limit}% used)

### Alerts
{For each alert in response.alerts}
- [{severity}] {message}
  Recommendation: {recommendation}
{End for}

### Cost by Model
| Model | Cost | Input Tokens | Output Tokens | Cache Read |
|-------|------|--------------|---------------|------------|
{For each model in cost_by_model}
| {model.model} | ${model.cost} | {model.input_tokens:,} | {model.output_tokens:,} | {model.cache_read_tokens:,} |
{End for}

### Token Breakdown
- Input: {tokens_by_type.input:,}
- Output: {tokens_by_type.output:,}
- Cache Read: {tokens_by_type.cacheRead:,}
- Cache Creation: {tokens_by_type.cacheCreation:,}

### Productivity Metrics
- Sessions: {productivity.sessions}
- Active Time: {productivity.active_time_formatted}
- Commits: {productivity.commits}
- PRs Created: {productivity.pull_requests}
- Lines Added: {productivity.lines_added:,}
- Lines Removed: {productivity.lines_removed:,}

### Notes
{Include any critical/warning alerts or observations}

When OTel is Not Enabled

## Cost Report Unavailable

OpenTelemetry metrics are not enabled on this platform.

{Include the setup_instructions from the response}

See docs/DEPLOYMENT.md for detailed setup instructions.