[api-consumption] 📊 GitHub API Consumption Report — 2026-08-05

[github-actions[bot]]

📊 GitHub API Consumption Report

Report Date: 2026-08-05 · Repository: github/gh-aw · Run: #31001504443

---

Today at a Glance

Metric	Value
🤖 Total Runs	140 (126 ✅ / 14 ❌)
🎯 Success Rate	90.0%
🔗 GitHub REST API Calls	102,161 (core quota consumed — includes reads, writes, and all GitHub API operations; partial day, still in progress)
📝 Safe-Output Writes	153 (issues + PRs + comments + discussions created by safe-output tools)
⏱ Avg Duration	623.6s (p95: 1316.0s)

---

🔗 GitHub API Calls Trend (90 days)

Only 3 days of history are available (2026-08-03 → 2026-08-05) — this repo runs such a high volume of agentic workflows (~150–210 runs/day) that a bounded log-fetch budget can't reach further back; see the cache status note below. Within that window, calls peaked on 08-04 at 249,489 (213 runs) before easing to 102,161 so far today (140 runs, day still in progress). No spikes beyond that single peak day are visible yet.

---

🔗 GitHub API Calls by Workflow Trend (30 days)

Across the available 3-day window, PR Sous Chef, Test Quality Sentinel, Matt Pocock Skills Reviewer, Impeccable Skills Reviewer, and Design Decision Gate consistently rank as the top consumers each day — the same cluster of PR-review/skills-check workflows drives most of the quota usage. "Other" (the remaining ~80+ lower-volume workflows combined) still outweighs any single top workflow, indicating consumption is broadly distributed rather than dominated by one script.

---

🔗 GitHub REST API Calls Heatmap (90 days)

The heatmap covers the same 2 raw-data days (08-04, 08-05, UTC hours). The busiest single hour was 20:00 UTC on 08-04 at 78,034 calls — a burst likely from several scheduled workflows firing in the same hour window rather than one workflow alone. Activity is otherwise spread fairly evenly across daytime/evening UTC hours with a quieter early-morning trough.

---

🍩 Top API Burners (24h)

No single workflow dominates today: the top 3 (Test Quality Sentinel, PR Sous Chef, Issue Monster) each account for roughly 7% of the day's total, and the remaining 63 lower-volume workflows combine for about 63% of consumption. This is a healthy, low-concentration-risk distribution — no one workflow is a single point of quota exhaustion.

---

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

Test Quality Sentinel (7,493 calls / 5 runs) and PR Sous Chef (7,463 calls / 8 runs) lead today's consumption, followed by Issue Monster (5,489 / 4 runs). None of the top 10 workflows are individually close to the 15,000/hr per-token limit even at their peak run, but PR Code Quality Reviewer (1048.7s avg) and Avenger (938.4s avg) have notably long average durations worth watching if run frequency increases.

---

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

Workflow	REST API Calls	Runs	Avg Duration
Test Quality Sentinel	7,493	5	484.4s
PR Sous Chef	7,463	8	473.4s
Issue Monster	5,489	4	388.0s
Impeccable Skills Reviewer	4,592	3	470.0s
Matt Pocock Skills Reviewer	4,584	3	428.7s
PR Description Updater	4,288	3	381.0s
Daily GitHub Docs SEO Optimizer	3,444	2	447.5s
Design Decision Gate 🏗️	3,061	3	415.3s
Avenger	3,005	8	938.4s
PR Code Quality Reviewer	2,923	3	1048.7s

---

Trending Indicators

• 7-day API trend: → insufficient data (only 3 days of history exist; see cache status)
• 30-day API trend: → insufficient data (only 3 days of history exist; see cache status)
• GitHub REST API call rate: ~152,530 calls/day averaged over the available 3-day window (hourly limit: 15,000 per token) — the busiest single hour (78,034 calls on 08-04 20:00 UTC) reflects aggregate consumption across many concurrently-scheduled workflows, each typically using its own token/app quota, not a single shared bucket

---

📦 Cache Memory Status

• Location: /tmp/gh-aw/cache-memory/trending/api-consumption/history.jsonl
• Cache restored from previous run: yes (2 existing entries: 2026-08-03, 2026-08-04)
• Collection mode: backfill
• Logs start_date used: -90d (requested), but this repo's workflow run volume (~150–210 runs/day) meant the 3-call bounded continuation budget (300 runs total) only reached back to 2026-08-04 in practice
• Data points stored: 3
• Earliest entry: 2026-08-03
• Retention policy: 90 days
• Note: a freshly re-fetched sample for 2026-08-04 (165 runs / 244,938 calls) was less complete than the already-cached entry for that date (213 runs / 249,489 calls, captured mid-day). To avoid regressing history with a less-complete resample, the existing cached 08-04 entry was kept as-is and only today's (2026-08-05) entry was added/updated.

---

Automatically generated by the api-consumption-report workflow.

Generated by 📊 GitHub API Consumption Report Agent · agent · 254.4 AIC · ⌖ 34.1 AIC · ⊞ 6.9K · ◷

• expires on Aug 8, 2026, 3:55 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 #50833.