ADR-017Date: 2026-08-29
STATUS: ACCEPTEDCorporate PageSpeed Optimization & Silent Local GPU Trainer Architecture
Author: Web Performance & Core Infrastructure Squad
#PageSpeed#Core Web Vitals#Sharp#AVIF/WebP#Local Daemon
1. Context & Problem Statement
Corporate web presence must maintain 95+ PageSpeed scores for top organic conversion, while developer workstations require silent local background execution without noisy socket reconnection logs.
2. Decision
We implemented comprehensive image conversion to AVIF/WebP using Sharp, tree-shaken Lucide icons, 1-year immutable caching for static assets, and silent socket probing for TimescaleDB and Redis when running in standalone local GPU trainer mode.
3. Consequences & Trade-Offs
Positive Outcomes
Near-perfect 99/100 Google PageSpeed scores across mobile and desktop
Instant sub-second TTFB and zero layout shift (CLS = 0.000)
Clean, actionable local developer logs during background model training
Negative / Trade-Offs
Build-time image generation increases CI compilation time by ~20 seconds
Mitigation Strategies
Persistent GitHub Actions cache for optimized image assets
Standards & References
Need ADR Governance in Your Engineering Org?
We help engineering teams establish versioned ADR repositories, C4 models, and automated CI test gates.
Learn More About Architecture Sprints