mcpField
M
magento-cache-analyst

magento-cache-analyst

github.com/maxnormDevelopment31143 installsUpdated 40w ago
Repository

About magento-cache-analyst

A cache analysis skill for Magento 2 developers, covering cache strategy, configuration, invalidation, and performance troubleshooting. It addresses full-page and application caches as well as distributed caching with Redis and Memcached.

Key features

  1. 01Full-page cache and ESI configuration
  2. 02Block, configuration, and layout caching
  3. 03Redis and Memcached configuration
  4. 04Cache invalidation and warming strategies
  5. 05Cache performance monitoring and troubleshooting

Use cases

  1. 01Analyze Magento cache bottlenecks
  2. 02Plan layered caching and storage backends
  3. 03Design cache clearing and synchronization strategies

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/maxnorm/magento2-agent-skills --skill magento-cache-analyst

After installing, type/magento-cache-analystin Claude Code to confirm it loaded.

Details

From the project README

Magento 2 Cache Analyst

Expert specialist in designing and implementing comprehensive caching strategies that dramatically improve application performance while ensuring data consistency and cache coherence across enterprise environments.

When to Use

  • Optimizing cache performance
  • Configuring full-page cache (FPC)
  • Implementing Redis or Memcached
  • Designing cache invalidation strategies
  • Troubleshooting cache issues
  • Planning cache architecture

Magento Cache Architecture

  • Cache Types: Master all Magento cache types and their optimal usage patterns
  • Full Page Cache: Expert in FPC configuration, ESI holes, and cache warming
  • Application Cache: Optimize block cache, configuration cache, and layout cache
  • Distributed Caching: Implement Redis, Memcached, and multi-tier caching
  • Cache Invalidation: Design efficient cache invalidation and purging strategies

Cache Types

Full Page Cache (FPC)
  • Varnish Configuration: Expert Varnish configuration and VCL optimization
  • Built-in FPC: Optimize Magento's built-in full-page cache
  • ESI Implementation: Implement Edge Side Includes for dynamic content
  • Cache Warming: Automated cache warming and preloading strategies
  • Cache Tags: Proper cache tag implementation for invalidation
Application Cache
  • Block Cache: Cache rendered blocks for faster page generation
  • Configuration Cache: Cache system and module configurations
  • Layout Cache: Cache layout XML structures
  • Collection Cache: Cache frequently accessed collections
  • Object Cache: Cache expensive object instantiation
Distributed Caching
  • Redis: High-performance in-memory data store
  • Memcached: Distributed memory caching system
  • Multi-tier Caching: Layer caching for optimal performance
  • Cache Synchronization: Maintain cache consistency across servers
  • Failover Strategies: Implement cache failover and recovery

Cache Optimization Process

1. Cache Assessment & Analysis
  • Performance Baseline: Establish current cache performance metrics
  • Cache Audit: Analyze existing cache configuration and effectiveness
  • Bottleneck Identification: Identify cache-related performance bottlenecks
  • Usage Pattern Analysis: Understand application cache usage patterns
  • Capacity Planning: Plan cache capacity and resource requirements
2. Cache Strategy Design
  • Layered Caching: Design multi-tier cache architectures
  • Cache Policies: Define cache TTL, eviction, and invalidation policies
  • Data Segmentation: Segment cache data for optimal performance
  • Storage Strategy: Choose appropriate cache storage backends
  • Synchronization Planning: Plan cache synchronization across environments
3. Implementation & Configuration
Redis Configuration
// app/etc/env.php
'cache' => [
    'frontend' => [
        'default' => [
            'backend' => 'Cm_Cache_Backend_Redis',
            'backend_options' => [
                'server' => '127.0.0.1',
                'port' => '6379',
                'database' => '0',
            ]
        ]
    ]
]
Cache Commands
# Clear all cache
bin/magento cache:clean
bin/magento cache:flush

# Enable/disable cache types
bin/magento cache:enable
bin/magento cache:disable

# Cache status
bin/magento cache:status
4. Testing & Optimization
  • Performance Testing: Validate cache performance improvements
  • Load Testing: Test cache behavior under high traffic conditions
  • Cache Warming: Implement effective cache warming strategies
  • Invalidation Testing: Test cache invalidation scenarios
  • Monitoring Validation: Verify monitoring and alerting effectiveness

Best Practices

Cache Strategy
  • Cache Hit Ratios: Maximize cache effectiveness and hit ratios
  • Memory Management: Optimize cache memory usage and allocation
  • Network Optimization: Reduce cache-related network overhead
  • Storage Optimization: Optimize cache storage and persistence strategies
  • Monitoring Integration: Implement comprehensive cache performance monitoring
Cache Invalidation
  • Tag-based Invalidation: Use cache tags for efficient invalidation
  • Event-based Invalidation: Invalidate cache on relevant events
  • Time-based Expiration: Set appropriate TTL values
  • Manual Invalidation: Provide manual cache clearing mechanisms
  • Selective Invalidation: Invalidate only affected cache entries
Enterprise Caching
  • CDN Integration: Optimize CDN caching and edge delivery
  • Multi-server Environments: Design caching for clustered deployments
  • Cache Synchronization: Maintain cache consistency across servers
  • Failover Strategies: Implement cache failover and recovery mechanisms
  • Scalability Planning: Design caching for horizontal and vertical scaling

Monitoring

  • Cache Hit Rates: Monitor cache hit/miss ratios
  • Memory Usage: Monitor cache memory consumption
  • Performance Metrics: Track cache-related performance improvements
  • Error Monitoring: Monitor cache errors and failures
  • Alerting: Set up alerts for cache issues

References

Focus on creating comprehensive caching strategies that dramatically improve performance while maintaining data consistency.

Comments

Markdown supported