Problem
The [aw] * failed failure reporter workflows have no deduplication logic. On May 26, 2026 alone, 20 failure issues were created — with Smoke Antigravity filed 3× in ~2 hours, Avenger 2×, and Daily News 2×. Over 60% of these are duplicates of existing open issues.
Root Cause
Each retry of a failing workflow creates a new issue without checking whether an open issue for the same workflow failure already exists.
Proposed Fix
Before creating a [aw] <workflow-name> failed issue, check for an existing open issue with the same title pattern. If one exists, add a comment to the existing issue instead of creating a new one.
Approximately:
1. Query open issues with label `agentic-workflows` and title matching `[aw] <workflow-name> failed`
2. If found: add comment with new run ID + timestamp
3. If not found: create new issue as normal
Expected Impact
- Reduce
[aw] failed issue volume from ~20/day to ~6-8/day (~60-70% reduction)
- Reduce issue list noise and improve signal quality
- Reduce reviewer fatigue from duplicate failure notifications
Context
Identified during Agent Performance Analysis run §26451465997.
Similar deduplication logic may already exist in Issue Monster — this could be extended to cover failure reporters.
Labels
agentic-workflows
optimization
automation
Generated by ⚡ Agent Performance Analyzer - Meta-Orchestrator · sonnet46 2.4M · ◷
Problem
The
[aw] * failedfailure reporter workflows have no deduplication logic. On May 26, 2026 alone, 20 failure issues were created — with Smoke Antigravity filed 3× in ~2 hours, Avenger 2×, and Daily News 2×. Over 60% of these are duplicates of existing open issues.Root Cause
Each retry of a failing workflow creates a new issue without checking whether an open issue for the same workflow failure already exists.
Proposed Fix
Before creating a
[aw] <workflow-name> failedissue, check for an existing open issue with the same title pattern. If one exists, add a comment to the existing issue instead of creating a new one.Approximately:
Expected Impact
[aw] failedissue volume from ~20/day to ~6-8/day (~60-70% reduction)Context
Identified during Agent Performance Analysis run §26451465997.
Similar deduplication logic may already exist in Issue Monster — this could be extended to cover failure reporters.
Labels
agentic-workflowsoptimizationautomation