+

Search Tips   |   Advanced Search

Tune dynamic cache with the cache monitor

Use this task to interpret cache monitor statistics to improve the performance of the dynamic cache service.

Use the cache monitor to watch cache hits versus misses. By comparing these two values, we can determine how much dynamic cache is helping the application, and if we can take any additional steps to further improve performance and decrease the cost of processing for the application server.


Tasks

  1. Start cache monitor and click on Cache Statistics. We can view the following cache statistics:

    Cache statistic Description
    Cache Size The maximum number of entries the cache can hold.
    Used Entries Number of cache entries used.
    Cache Hits Number of request responses served from the cache.
    Cache Misses Number of request responses that are cacheable but cannot be served from the cache.
    LRU Evictions Number of cache entries removed to make room for new cache entries.
    Explicit Removals Number of cache entries removed or invalidated from the cache based on cache policies or were deleted from the cache through the cache monitor.

  2. We can also view the following cache configuration values:

    Cache configuration value Description
    Default priority Lower priority entries are moved from the cache before higher priority entries when the cache is full. We can specify the priority for individual cache entries in the cache policy.
    Servlet Caching Enabled If servlet caching is enabled, results from servlets and JSP files are cached.
    Disk Offload Enabled Entries that are being removed from the cache are saved to disk.

  3. Wait for the application server to add data to the cache.

    We want the number of used cache entries in the cache monitor to be as high as it can go. When the number of used entries is at its highest, the cache can serve responses to as many requests as possible.

  4. When the cache has a high number of used entries, reset the statistics.

    Watch the number of cache hits versus cache misses. If the number of hits is far greater than the number of misses, the cache configuration is optimal. We do not need to take any further actions. If we find a higher number of misses with a lower number of hits, the application server is working hard to generate responses instead of serving the request using a cached value. The application server might be making database queries, or running logic to respond to the requests.

  5. If we have a large number of cache misses, increase the number of cache hits by improving the probability that a request can be served from the cache.

    Increase the cache size or configure additional cache policies.


Related:

  • Cache monitor
  • Troubleshoot the dynamic cache service
  • Configure servlet caching
  • Configure dynamic cache disk offload
  • Configure cacheable objects with the cachespec.xml file
  • Task overview: Using the dynamic cache service to improve performance
  • Dynamic cache service settings