cache-diagnostic_skill

Overview

This skill performs a focused Magento 2 cache diagnostic to identify configuration issues, measure performance, and surface optimization opportunities across all cache layers. It evaluates Full Page Cache, application caches, Redis, and Varnish to deliver actionable recommendations that reduce latency and improve hit rates. The output highlights misconfigurations, resource pressure, and quick wins for cache stability and speed.

How this skill works

The skill inspects Magento cache status, backend settings, and lifetime parameters, then collects live metrics from Redis and Varnish when present. It computes hit/miss ratios, analyzes Redis memory usage and slow logs, and validates VCL and backend health. Results combine configuration checks, performance metrics, and tailored optimization steps prioritized by impact.

When to use it

• After a site migration or environment change that affects cache backends
• When page load times or TTFB increase without clear cause
• If Redis shows memory pressure, eviction, or unexplained latency
• When Varnish cache hit rates are low or PURGE/invalidations misbehave
• As part of routine performance audits before peak traffic events

Best practices

• Ensure a single authoritative FPC backend is configured and avoid mixed backends across environments
• Use Redis with appropriate maxmemory and eviction policy for session and cache segregation
• Monitor Redis keyspace and fragmentation; tune persistence and eviction to avoid OOM events
• Validate Varnish VCL for correct ESI and backend health checks; automate VCL deployment safely
• Warm caches after deployment and reduce unnecessary invalidation by targeting tags and lifetimes

Example use cases

• Detecting misconfigured cache backend after switching environments (file vs Redis)
• Investigating low FPC hit rate and recommending VCL or cacheability adjustments
• Identifying Redis memory fragmentation, recommending maxmemory and reindex strategies
• Validating Varnish configuration and backend response behavior during traffic spikes
• Prioritizing cache-related fixes before a planned marketing campaign to avoid cache stampedes