[api-consumption] 📊 GitHub API Consumption Report — 2026-08-28

[github-actions[bot]]

📊 GitHub API Consumption Report

Report Date: 2026-08-28 · Repository: github/gh-aw · Run: #33181002249

---

Today at a Glance

Metric	Value
🤖 Total Runs	252 (227 ✅ / 25 ❌)
🎯 Success Rate	90.08%
🔗 GitHub REST API Calls	96,822 (core quota consumed — includes reads, writes, and all GitHub API operations)
📝 Safe-Output Writes	284 (issues + PRs + comments + discussions created by safe-output tools)
⏱ Avg Duration	586.02s (p95: 1179.0s)

⚠️ Data quality note: The logs MCP tool returned partial results at every pagination step. After 2 continuation calls (the policy limit), the collected dataset covers only 2026-08-27T23:14Z through 2026-08-28T14:14Z (~15 hours), not the requested 90-day backfill window. All "last 24h" figures above reflect this partial window. Additionally, 80 of 333 downloaded run directories had no run_summary.json yet (likely still processing) and were excluded from aggregation.

---

🔗 GitHub API Calls Trend (90 days)

The 10 data points on record (spanning 2026-08-14 to 2026-08-28, with several gaps where no report ran) show high day-to-day volatility rather than a steady trend: a peak of 398,195 calls on 2026-08-15, a trough of 245 calls on 2026-08-24, and another peak of 290,274 on 2026-08-22. Today's partial-window figure of 96,822 sits mid-range, but given the ~15-hour collection gap it likely understates the full day's consumption.

---

🔗 GitHub API Calls by Workflow Trend (30 days)

Review-style workflows dominate: PR Sous Chef, Impeccable Skills Reviewer, PR Code Quality Reviewer, Matt Pocock Skills Reviewer, and Ponytail Reviewer appear repeatedly among the top consumers across the days we have data for. These are recurring PR-review agents that run once per PR event, so their footprint scales directly with PR volume rather than schedule.

---

🔗 GitHub REST API Calls Heatmap (90 days)

With only 10 sparse days on record (several calendar days have no entry at all — reporting gaps, not necessarily zero activity), the heatmap is too sparse to identify a reliable weekly pattern yet. Continued daily collection will fill this in over the next few weeks.

---

🍩 Top API Burners (24h)

Within today's partial window, PR Description Updater (10,296 calls), Impeccable Skills Reviewer (8,721), and PR Code Quality Reviewer (8,209) together account for roughly 28% of total consumption. No single workflow dominates outright — the top 10 workflows are fairly evenly spread between ~3,000 and ~10,300 calls each, so concentration risk is moderate rather than severe.

---

🔗 GitHub REST API Consumption by Workflow (last 24h)

The heaviest consumers in the partial window are PR-review workflows (PR Description Updater, Impeccable Skills Reviewer, PR Code Quality Reviewer, Matt Pocock Skills Reviewer, Ponytail Reviewer, Design Decision Gate), each in the 6,800–10,300 call range across 11–15 runs — none is close to the 15,000/hr hourly limit individually. Copilot Agent Prompt Clustering Analysis stands out for a single run taking 1,872s (31 min) while consuming only 3,004 calls, suggesting its cost is duration-bound rather than API-bound; that's a candidate for reviewing whether its long runtime is expected.

---

Top 10 Workflows by REST API Consumption (last 24h)

Workflow	REST API Calls	Runs	Avg Duration
PR Description Updater	10,296	11	416.27s
Impeccable Skills Reviewer	8,721	15	451.47s
PR Code Quality Reviewer	8,209	15	405.73s
Matt Pocock Skills Reviewer	7,736	15	466.93s
Ponytail Reviewer	7,611	14	394.79s
Design Decision Gate 🏗️	6,807	14	427.79s
Daily Trajectory Grader Implementer	3,687	1	486.0s
Copilot Agent Prompt Clustering Analysis	3,004	1	1872.0s
Lockfile Statistics Analysis Agent	2,985	1	620.0s
Daily Team Evolution Insights	2,976	1	1194.0s

---

Trending Indicators

• 7-day API trend: → insufficient data (only 10 sparse data points on record; need ≥14 for a reliable 7-day-over-7-day comparison)
• 30-day API trend: → insufficient data (need ≥60 data points; only 10 on record)
• GitHub REST API call rate: ~126,306 calls/day average over the last 7 recorded data points (hourly limit: 15,000) — note this average is skewed by sparse, uneven daily coverage rather than a true 7-consecutive-day rate

---

📦 Cache Memory Status

• Location: /tmp/gh-aw/cache-memory/trending/api-consumption/history.jsonl
• Cache restored from previous run: yes (9 existing entries)
• Collection mode: backfill
• Logs start_date used: -90d (2 continuation calls made, policy limit reached; effective coverage ~15 hours)
• Data points stored: 10
• Earliest entry: 2026-08-14
• Retention policy: 90 days

---

Automatically generated by the api-consumption-report workflow.

Warning

Firewall blocked 1 domain

The following domain was blocked by the firewall during workflow execution:

• api.anthropic.com

To allow these domains, add them to the network.allowed list in your workflow frontmatter:

network:
  allowed:
    - defaults
    - "api.anthropic.com"

See Network Configuration for more information.

Generated by 📊 GitHub API Consumption Report Agent · claude · agent · 122.7 AIC · ⌖ 9.48 AIC · ⊞ 13.4K · ◷

• expires on Aug 31, 2026, 6:54 AM UTC-08:00

[github-actions[bot]]

This discussion has been marked as outdated by GitHub API Consumption Report Agent.

A newer discussion is available at Discussion #56875.