[api-consumption] 📊 GitHub API Consumption Report — 2026-08-07

[github-actions[bot]]

📊 GitHub API Consumption Report

Report Date: 2026-08-07 · Repository: github/gh-aw · Run: #31172392612

---

Today at a Glance

Metric	Value
🤖 Total Runs	299 (214 ✅ / 85 ❌)
🎯 Success Rate	71.6%
🔗 GitHub REST API Calls	4,506 (core quota consumed — includes reads, writes, and all GitHub API operations)
📝 Safe-Output Writes	not available this run (see Cache Memory Status note)
⏱ Avg Duration	2372.0s (p95: 12708.0s)

---

🔗 GitHub API Calls Trend (90 days)

This is the first run of the trending report, so only two calendar days of history are plotted: 2026-08-06 (1,638 calls, partial — data starts at 14:49 UTC) and 2026-08-07 (2,868 calls, partial — data ends at the time of this run, 10:51 UTC). No 90-day trend can be computed yet; the apparent rise reflects partial-day coverage, not necessarily a real trend. Coverage will fill in as this workflow runs daily.

---

🔗 GitHub API Calls by Workflow Trend (30 days)

With only two partial days on record, per-workflow movement (e.g. PR Description Updater climbing from 50→353 calls, AI Moderator dropping from 303→91) mostly reflects which hours of each day were captured rather than a genuine multi-day trend. Impeccable Skills Reviewer and PR Sous Chef are the most consistent heavy consumers across both partial days.

---

🔗 GitHub REST API Calls Heatmap (90 days)

Only Thursday (2026-08-06, hours 14–23 UTC) and Friday (2026-08-07, hours 00–10 UTC) have data — together they form one continuous ~20-hour window, not two independent days of pattern. Within that window, the busiest stretch is 00:00–06:00 UTC on Friday (several 300+ call hours, peaking near 650 calls at 23:00), while the Thursday afternoon/evening hours run lighter (mostly under 150 calls/hour). No weekly pattern can be inferred yet from a single day-and-a-half of data.

---

🍩 Top API Burners (24h)

The top 7 named workflows account for 49% of the 4,506 calls collected, led by Impeccable Skills Reviewer (12%) and AI Moderator (9%); the remaining 51% ("Other") is spread thinly across dozens of lower-volume workflows in this very active repository, so concentration risk in any single workflow is low.

---

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

Impeccable Skills Reviewer (525 calls / 18 runs, ~29/run) and AI Moderator (394 calls / 5 runs, ~79/run) are the top two consumers; AI Moderator's high per-run average is the standout efficiency outlier worth a closer look. All workflows are far below the 15,000/hr core rate limit — no single run or workflow is close to throttling risk in this window.

---

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

Workflow	REST API Calls	Runs	Avg Duration
Impeccable Skills Reviewer	525	18	2667s
AI Moderator	394	5	2605s
PR Description Updater	353	8	2480s
PR Sous Chef	306	25	2563s
Test Quality Sentinel	232	21	3556s
PR Code Quality Reviewer	194	18	2091s
Design Decision Gate	189	19	2675s
Issue Monster	185	14	2844s
Matt Pocock Skills Reviewer	157	20	2337s
Auto-Triage Issues	98	6	3537s

---

Trending Indicators

• 7-day API trend: → insufficient data (only 2 partial days recorded — first run)
• 30-day API trend: → insufficient data (only 2 partial days recorded — first run)
• GitHub REST API call rate: ~4,506 calls in ~20h of collected history (hourly limit: 15,000) — well within headroom

---

⚠️ Notable observations

• Run health dipped on 2026-08-06: only 43.9% success (58/132) vs. 93.4% (156/167) on 2026-08-07. Failures on 08-06 were concentrated between 15:00–19:00 UTC (52 of 74 failures) and spread across many different workflows (top: PR Code Quality Reviewer 7, Design Decision Gate 5, Issue Monster 5, PR Sous Chef 5) rather than one dominant cause — worth a look if this recurs.
• Safe-output write counts are not included this run. The MCP logs tool's default artifact scope only downloads the usage artifact; safe_output.jsonl/run_summary.json's safe-output fields require --artifacts all (or agent), which was out of scope for the bounded collection budget used here.

---

📦 Cache Memory Status

• Location: /tmp/gh-aw/cache-memory/trending/api-consumption/history.jsonl
• Cache restored from previous run: no (first run — cache was empty)
• Collection mode: backfill (target: 90 days)
• Logs start_date used: -90d, with 2 additional continuation calls (as permitted)
• Actual coverage achieved: ~20 hours (2026-08-06 14:49 UTC → 2026-08-07 10:51 UTC), 299 unique runs across 3 log-fetch calls. This repository produces a very high volume of workflow runs (150+/day), so the 100-runs-per-call × 3-calls budget was exhausted well before reaching 90 days back. This is expected and not a tool failure — coverage will accumulate incrementally on each future daily run until the 90-day window is filled.
• Data points stored: 2 (both partial-day)
• Earliest entry: 2026-08-06
• Retention policy: 90 days

---

Automatically generated by the api-consumption-report workflow.

Generated by 📊 GitHub API Consumption Report Agent · agent · 148.3 AIC · ⊞ 6.9K · ◷

• expires on Aug 10, 2026, 3:18 AM UTC-08:00