NCache 4.4 - Online Documentation

Performance Counters

 
NCache is not only enriched with powerful monitoring facilities, but besides that; it also provides different categories of
counters for windows performance monitor. This lets the clients monitor the overall performance of NCache. These
categories include cluster, client, server and bridge. They cover all aspects of NCache monitoring including remote
clients, cache clients, clustered operations, server counters, bridge queue status counters and many more.
 
Performance counters provide valuable information about system hardware, services, networks and applications that can be used to determine their performance.  It helps in diagnosing problems, used to fine-tune systems and
applications, monitoring real-time application performance and resolving bottleneck cause of system components. 
 
Performance counters can be seen both from Perfmon and NCache Monitor. The statistics view is also used to see the performance statistics of a cache or a cluster.
 
These counters include:
 
  • Fetches/sec: Number of Get operations per second.
  • Hits/sec: Number of successful Get operations per second.
  • Misses/sec: Number of failed Get operations per second.
  • Additions/sec: Number of Add operations per second.
  • Updates/sec: Number of Insert operations per second.
  • Deletes/sec: Number of Remove operations per second.
  • Expirations/sec: Number of items being expired currently per second.
  • Evictions/sec: Number of items evicted per second.
  • State Txfr/sec: Number of items this node is either reading from other nodes or sending to other nodes during a state transfer mode.
  • Data Bal/sec: Number of items this node is either reading from other nodes or sending to other nodes during a Data Load Balancing mode.
  • Mirror Queue Size: Number of items in the Mirror queue.
  • Average sec/fetch: Average time in seconds taken to complete one fetch operation.
  • Average sec/add: Average time in seconds taken to complete one Add operation.
  • Average sec/insert: Average time in seconds taken to complete one insert operation.
  • Average sec/remove: Average time in seconds taken to complete one remove operation.
  • Average sec/cache operation: Average time, in seconds, taken to complete one cache operation.
  • Client Request/sec: Number of requests sent by all clients to the cache server.
  • Client Response/sec: Number of responses received by all clients from the cache server.
  • Clustered opsrecv/sec: Number of clustered operations performed per second.
  • Clustered opsent/sec: Number of clustered operations performed per second.
  • Count: Number of items in the cache.
  • DispatchEnter/sec: Number of clustered operations performed per second.
  • Mirror queue size: Number of items in the Mirror queue.
  • Misses/sec: Number of failed Get operations per second.
  • Readthru/sec: Number of Readthru operations per second.
  • Response sent/sec: Number of clustered operations performed per second.
  • Responses/sec: Number of responses sent (meaning cache response for commands like add, get, insert, remove etc.) to all clients by this cache server.
  • State transfer/sec: Number of items this node is either reading from other nodes or sending to other nodes during a state transfer mode.
  • Writethru/sec: Number of Writethru/Writebehind operations per second.
  • CacheLastAccessCount: Number of items which are older then the access interval specified in the config.
  • Cluster ops/sec: Number of clustered operations performed per second.
  • Bytes sent/sec: How much data is sent from this node to other nodes in the cache.
  • Bytes received/sec: How much data is received by this node from other nodes in the cache.
  • Requests/sec: Number of requests (meaning cache commands like add, get, insert, remove etc.) being processed from all clients by this cache server.
  • Client bytes sent/sec: Bytes being sent from cache server to all its clients.
  • Client bytes received/sec: Bytes being received by cache server from all its clients.
  • Response Queue Count: Number of items in response queue; all responses are queued in the response queue before being sent.
  • Response Queue Size: Size of the response queue in bytes. Is displays the total data size of all the responses in the response queue.
  • Event Queue Count: Number of events in the event queue.
  • Events Processed/sec: Number of events processed per second on a client.
  • Events Triggered/sec: Number of events triggered and received by client per second.
  • Average ms/event: Average time taken by client to process a single event.
 
You can select different counters to be displayed in 'Cluster Performance Test Bed' to analyze cluster performance through Options Dialog. You can enable Performance Counters for both Local and Clustered Caches from Options Tab of their view window.
 
 
See Also
Debugging Counters | Monitoring Caches using NCache Manager | Monitoring Caches using NCache Monitor | Monitoring Bridge using NCache Manager | Monitoring NCache using Windows Perfmon Tool | Configuring Caches
Send comments on this topic.