[Parent] Track and resolve recurring AI-credits rate-limit failures in scheduled agentic workflows

[github-actions]

Scheduled agentic workflows (Issue Arborist, Dependency Version Updater, and others) are repeatedly hitting the AI-credits rate limit, causing failed or incomplete pipeline runs. These failures have accumulated into 7+ separate incident issues since they share a common root cause.

Related incident reports

• [aw] Issue Arborist hit AI credits rate limit #1515 — Issue Arborist hit AI credits rate limit
• [aw] Dependency Version Updater hit AI credits rate limit #1492 — Dependency Version Updater hit AI credits rate limit
• [aw] Dependency Version Updater hit AI credits rate limit #1478 — Dependency Version Updater hit AI credits rate limit
• [aw] Issue Arborist hit AI credits rate limit #1467 — Issue Arborist hit AI credits rate limit
• [aw] Dependency Version Updater hit AI credits rate limit #1441 — Dependency Version Updater hit AI credits rate limit
• [aw] Issue Arborist hit AI credits rate limit #1430 — Issue Arborist hit AI credits rate limit
• [aw] Dependency Version Updater hit AI credits rate limit #1412 — Dependency Version Updater hit AI credits rate limit

Potential mitigations

• Reduce token usage per agent run (tighter prompts, shorter context)
• Stagger or reduce schedule frequency
• Investigate whether model selection affects credit consumption
• Add retry/backoff strategy when rate-limited rather than failing immediately
• Surface credit-usage telemetry in ado-aw audit for ongoing monitoring

Generated by Issue Arborist · 44.2 AIC · ⌖ 6.15 AIC · ⊞ 6K · ◷