[api-consumption] 📊 GitHub API Consumption Report — 2026-08-02

[github-actions[bot]]

📊 GitHub API Consumption Report

Report Date: 2026-08-02 · Repository: github/gh-aw · Run: #30745112201

---

Today at a Glance

Metric	Value
🤖 Total Runs	306 (291 ✅ / 15 ❌)
🎯 Success Rate	95.1%
🔗 GitHub REST API Calls	250,500 (core quota consumed — includes reads, writes, and all GitHub API operations)
📝 Safe-Output Writes	0 (issues + PRs + comments + discussions created by safe-output tools)
⏱ Avg Duration	597s (p95: 1158s)

---

🔗 GitHub API Calls Trend (90 days)

Daily REST API consumption has trended upward over the 72-day window, with a peak of ~574K calls on 2026-06-17. The 7-day rolling average sits well above the window's early baseline, and today's total (250.5K) remains elevated relative to the trailing average — consistent with the growing number of active agentic workflows in this repository.

---

🔗 GitHub API Calls by Workflow Trend (30 days)

Over the last 30 days, a small cluster of PR-review and skills-reviewer style workflows (e.g. Matt Pocock Skills Reviewer, Impeccable Skills Reviewer, PR Code Quality Reviewer) consistently rank among the top consumers, reflecting their per-PR-event trigger pattern. No single workflow dominates the 30-day trend line — consumption is spread across several recurring PR-triggered agents rather than concentrated in one.

---

🔗 GitHub REST API Calls Heatmap (90 days)

The heatmap shows generally higher weekday consumption versus weekends, consistent with PR-driven and scheduled daily-report workflows running on a business-day cadence. The darkest cell (2026-06-17) stands out as a clear anomaly against an otherwise moderate, steadily-rising background.

---

🍩 Top API Burners (24h)

The top 10 workflows account for the large majority of today's REST API footprint, led by Matt Pocock Skills Reviewer (28.1K) and Impeccable Skills Reviewer (26.4K). Concentration risk is moderate: no single workflow exceeds ~11% of the 24h total, but the top 10 combined represent a substantial share versus the long tail of ~130 other workflows that ran today.

---

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

No individual workflow run is anywhere near the 15,000/hr hourly rate-limit line — the highest single-workflow total (28.1K) is aggregated across 17 runs, not one run. The main optimisation opportunity is in the PR-review family of workflows (Skills Reviewer variants, PR Code Quality Reviewer, PR Description Updater), which collectively account for a large share of daily consumption and could benefit from caching or reduced per-PR API calls.

---

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

Workflow	REST API Calls	Runs	Avg Duration
Matt Pocock Skills Reviewer	28,087	17	465s
Impeccable Skills Reviewer	26,385	16	438s
PR Code Quality Reviewer	23,543	16	686s
Design Decision Gate 🏗️	18,555	16	465s
PR Description Updater	15,818	13	426s
PR Sous Chef	14,974	35	532s
Test Quality Sentinel	12,549	17	530s
Smoke Copilot - AOAI (apikey)	7,273	2	884s
Avenger	7,229	8	840s
Smoke OTEL	6,130	2	548s

---

Trending Indicators

• 7-day API trend: ↑ +23.6% vs. previous 7 days
• 30-day API trend: ↑ +70.2% vs. prior 30 days
• GitHub REST API call rate: ~132,805 calls/day over last 7 days (hourly limit: 15,000)

---

📦 Cache Memory Status

• Location: /tmp/gh-aw/cache-memory/trending/api-consumption/history.jsonl
• Cache restored from previous run: yes (71 entries)
• Collection mode: incremental
• Logs start_date used: -1d
• Data points stored: 72
• Earliest entry: 2026-05-18
• Retention policy: 90 days

⚠️ Data collection notes

• Log collection reached the continuation limit twice; per policy, paging stopped after 2 continuation calls (300 total runs fetched across all pages) once continuation returned null. Of the ~597 run directories present on disk, 306 contained usable aw_info.json/run_summary.json data and were used for these metrics; the remainder were empty/stale artifacts and were excluded.
• 15 runs failed in the collection window (various workflows — PR Sous Chef, PR Code Quality Reviewer, Discussion Task Miner, Code Scanning Fixer, Smoke Copilot, Daily Credit Limit Test, Daily Container Image Security Scan, daily-experiment-report). These were not individually audited in this report since it focuses on API consumption rather than failure root-cause analysis.
• github_safe_output_calls is 0 for today because run.safe_items_count was null across all 306 collected run_summary.json files in this window.

---

Automatically generated by the api-consumption-report workflow.

Generated by 📊 GitHub API Consumption Report Agent · agent · 244.7 AIC · ⌖ 37.2 AIC · ⊞ 12.4K · ◷

• expires on Aug 5, 2026, 3:37 AM UTC-08:00