Skip to content

ECO-AIM-RAG-001

Name: Embedding regeneration without change detection

Category: AI/ML

Family: Retrieval-Augmented Generation

Primary layer: ai

System layers: ai

Description

Embeddings are regenerated for unchanged content, wasting compute and increasing pipeline latency.

Impact

  • type: ai-compute
  • confidence: 0.75
  • notes: Added as part of the 0.3.0 expansion to capture cross-system sustainability and operational waste.

Detection

  • method: static-or-runtime
  • confidence: 0.55
  • runtime_validation_required: Yes

Remediation

  • guidance: Use content hashing, incremental indexing, and idempotent embedding pipelines.
  • tradeoffs: May require architecture, product, or operations review rather than a local code change.

Cost Dimensions

  • compute: high
  • memory: medium
  • network: medium
  • storage: medium
  • human_time: low
  • carbon: high
  • water: medium

Amplification

  • scales_with_users: No
  • scales_with_data_volume: Yes
  • scales_non_linearly: No

Temporal Behavior

  • startup_only: No
  • steady_state: Yes
  • burst_sensitive: Yes
  • time_degradation: No

Runtime Evidence

  • embedding job logs
  • content hashes
  • vector DB writes

Pattern examples

No pattern examples provided.

Remediation examples

No remediation examples provided.

Metadata

  • catalog_version: 0.4.0
  • status: draft
  • source: catalog expansion recommendations applied 2026-05-21