Skip to content

[deep-report] Exclude intentional credit-guardrail tests from prod-main health rollups #47099

Description

@github-actions

Description

The Daily Credit Limit Test and Daily Max Ai Credits Test workflows are designed to fail (they trip the AI-credit guardrail on purpose). The audit-workflows report (discussion #46930) shows they depress the prod-main success rate by ~1.5 points every day (92.7% raw → 94.2% excluding them). Tag these workflows (e.g. an intentional-failure marker in frontmatter or an audit allow-list) and filter them out of the prod-main and fleet-health rollups.

Expected Impact

Health dashboards stop under-reporting fleet health by ~1.5pts daily, so real regressions become visible against a truer baseline.

Suggested Agent

Agentic Workflow Audit agent.

Estimated Effort

Quick (< 1 hour).

Data Source

DeepReport analysis 2026-07-21; discussion #46930 (recommendation #2).

Generated by 🔬 Deep Report · age00 200.9 AIC · ⌖ 17.2 AIC · ⊞ 9.9K ·

  • expires on Jul 23, 2026, 7:27 AM UTC-08:00

Metadata

Metadata

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions