📊 Daily Copilot Token Consumption Report - 2026-02-20 #17124
Replies: 1 comment
-
|
🤖 Smoke test agent was here! Beep boop! 👋 The Copilot smoke test agent just stopped by to say hello. All systems nominal, circuits humming, and tokens flowing freely. Carry on! 🚀
|
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
Executive Summary
Over the last 30 days (data window: Feb 19–20), Copilot-powered agentic workflows consumed 166,507,574 tokens across 195 workflow runs spanning 50 unique workflows. Token consumption is up +39.5% vs yesterday (Feb 19: 119M tokens), driven largely by the increase in run volume (+135% runs). Cost data is not available in current metrics.
Key Highlights
📈 Token Usage Trends
Overall Token & Run Trends (6 days of history)
Daily token consumption shows variability tied closely to run counts. Feb 11 saw the highest recorded day (237M tokens / 378 runs). Today's 166M tokens across 195 runs represents a moderate day.
🏆 Top Workflows by Token Consumption
Top 10 Most Expensive Workflows
Token Distribution
Per-Workflow Detailed Statistics (All 50 Workflows)
💡 Insights & Recommendations
High-Cost Workflows
The top 5 workflows account for 62.6% of all token consumption:
Chroma Issue Indexer — 25.9M tokens (15.6%) across 8 runs
CI Failure Doctor — 20.1M tokens (12.1%) across 8 runs
Example: Custom Error Patterns — 12.9M tokens (7.7%) across only 3 runs
Documentation Noob Tester — 10.9M tokens (6.6%) across 2 runs
Contribution Check — 7.6M tokens (4.5%) across 8 runs
Workflows with Zero Tokens (Errors)
Three workflows reported 0 tokens with errors, indicating they failed before producing output:
Issue Monster has a 100% failure rate (16/16 runs errored). This should be investigated as it wastes compute without producing useful output.
High-Volume Low-Cost Workflows (Efficient)
These workflows run frequently but consume tokens efficiently:
Smoke test workflows are well-optimized — they run the most frequently but stay under 210K tokens/run.
Historical Comparison
The highest-ever recorded day was Feb 11 (237M tokens / 378 runs). Today's consumption is 30% below that peak.
Methodology & Data Quality Notes
Methodology
/tmp/gh-aw/copilot-logs.json)memory/token-metricsbranch)Data Quality Notes
total_cost: 0in summary); cost estimates cannot be providedtotal_turnsin summary is reported as 4, which seems low — turn data may not be captured per-run in current log formatReferences:
Beta Was this translation helpful? Give feedback.
All reactions