📊 Copilot Token Usage Report2026-05-02

[github-actions]

Overview

Period: 2026-05-01T22:00Z to 2026-05-02T10:29Z (~12.5 hours)
Runs analyzed: 50 (38 had token data)
Total tokens: 5,390K (~5.4M) across all workflows
Estimated total cost: $0.00 (cost tracking via AWF api-proxy; USD cost not yet reported)
Total Actions minutes: 302 min
Total errors: 29 across all runs

Token consumption is down ~44% vs the previous report (9,542K on 2026-05-01), with 12 failed/cancelled runs producing zero tokens accounting for most of the reduction.

Workflow Summary

Sorted by total tokens (cost not available):

Workflow	Runs	Total Tokens	Avg Tokens/Run	Turns	Errors
Build Test Suite	8	1,442K	180K	4	6
Smoke Services	9	1,120K	124K	6	7
Smoke Copilot	11	986K	90K	8	6
Dependency Security Monitor	1	556K	556K	1	0
Smoke Copilot BYOK	9	491K	55K	5	5
Issue Duplication Detector	1	212K	212K	2	0
Firewall Issue Dispatcher	2	312K	156K	2	0
Smoke Chroot	8	272K	34K	4	5
Daily Copilot Token Usage Analyzer	1	0	—	0	0

🔍 Optimization Opportunities

1. Dependency Security Monitor — 556K tokens in a single run

   • Highest single-run token count; review whether prompt/context can be trimmed
   • Only runs once, so monitoring the trend is key

2. Issue Duplication Detector — 212K tokens/run, 2 turns

   • Relatively high for a single-purpose workflow; consider reducing the number of issues fetched per scan

3. Build Test Suite — avg 180K tokens/run

   • 5 of 8 runs had zero tokens (failures/cancellations early in the period); the successful runs average ~480K — review whether the test suite prompt can be scoped more narrowly

4. High error rate — Smoke Services (7/9 errors), Smoke Copilot (6/11), Smoke Copilot BYOK (5/9), Smoke Chroot (5/8), Build Test Suite (6/8)

   • Most errors appear clustered in runs between 23:49–00:26 UTC, suggesting an infrastructure incident rather than individual workflow issues

Per-Workflow Details

Build Test Suite

• Runs: 8
• Total tokens: 1,442K (avg 180K/run; successful runs avg ~480K)
• Turns: 4 total (avg 0.5/run)
• Error rate: 6/8 runs with errors
• Run links: 25239353589 ✅ 25238998766 ❌ 25238556894 ❌ 25238249639 ❌ 25238125747 ❌ 25236024273 ✅ 25235778695 ✅ 25235431476 🚫

Smoke Services

• Runs: 9
• Total tokens: 1,120K (avg 124K/run)
• Turns: 6 total
• Error rate: 7/9 runs with errors (most cancelled/failed)
• Run links: 25239967748 ✅ 25239353578 ✅ 25238998726 ✅ 25238556876 ❌ 25238249669 🚫 25238125722 🚫 25236024231 ✅ 25235778661 ✅ 25235431462 🚫

Smoke Copilot

• Runs: 11
• Total tokens: 986K (avg 90K/run)
• Turns: 8 total
• Error rate: 6/11 runs with errors
• Run links: 25239826547 ❌ 25239353566 ✅ 25238998722 ✅ 25238556917 ❌ 25238249684 ❌ 25238125729 ❌ 25236957203 ✅ 25236024251 ✅ 25235778646 ✅ 25235431530 🚫 25235013564 ✅

Dependency Security Monitor

• Runs: 1
• Total tokens: 556K
• Turns: 1
• Error rate: 0/1 runs
• Run link: 25237777048 ✅

Smoke Copilot BYOK

• Runs: 9
• Total tokens: 491K (avg 55K/run)
• Turns: 5 total
• Error rate: 5/9 runs with errors (4 failures with zero tokens)
• Run links: 25239996595 ✅ 25239353562 ✅ 25238998760 ❌ 25238556897 ❌ 25238249708 ❌ 25238125721 ❌ 25236024268 ✅ 25235778656 ✅ 25235431463 🚫

Firewall Issue Dispatcher

• Runs: 2
• Total tokens: 312K (avg 156K/run)
• Turns: 2 total
• Error rate: 0/2 runs
• Run links: 25246367284 ✅ 25239936066 ✅

Issue Duplication Detector

• Runs: 1
• Total tokens: 212K
• Turns: 2
• Error rate: 0/1 runs
• Run link: 25235155369 ✅

Smoke Chroot

• Runs: 8
• Total tokens: 272K (avg 34K/run)
• Turns: 4 total
• Error rate: 5/8 runs with errors
• Run links: 25239353560 ✅ 25238998745 ✅ 25238557070 ❌ 25238249645 ❌ 25238125728 ❌ 25236024242 ✅ 25235778696 ✅ 25235431470 🚫

Workflows Without Token Data

The following workflows had runs with token_usage = 0 or null — these are either early failures (before the agent started) or infrastructure-level cancellations:

• Daily Copilot Token Usage Analyzer — 1 run, currently in progress (this run)
• Smoke Copilot BYOK — 4 runs failed before agent started
• Smoke Copilot — 4 runs failed before agent started
• Smoke Services — 3 runs cancelled/failed before agent started
• Build Test Suite — 4 runs failed before agent started
• Smoke Chroot — 3 runs cancelled/failed before agent started

These are not instrumentation gaps — the failures occurred before any LLM inference was invoked.

Historical Trend

Date	Runs	Total Tokens	Notes
2026-04-29	—	—	Baseline
2026-04-30	—	~8,496K	Prior report
2026-05-01	50	~9,542K	+12% vs prev
2026-05-02	50	~5,390K	-44% vs prev

The sharp drop is primarily due to many failed/cancelled runs (29 errors, ~12 zero-token runs) between ~23:49–00:26 UTC suggesting an infrastructure incident. Successful run token counts are consistent with prior periods.

Previous Report

#2344 — 2026-05-01

References:

• §25249909632 (this run)
• §25239353589 (Build Test Suite top run)
• §25237777048 (Dependency Security Monitor)

Generated by Daily Copilot Token Usage Analyzer · ● 290.2K · ◷