6 Analyzing Reporter Content

Coherence provides out-of-box reports that helps administrators and developers better analyze usage and configuration issues that may occur. For reports specific to Coherence*Web, see "Running Performance Reports" in Oracle Coherence User's Guide for Oracle Coherence*Web.

The following sections are included in this chapter:

Understanding the Cache Size Report
Understanding the Cache Usage Report
Understanding the Flash Journal Report
Understanding the Management Report
Understanding the Memory Status Report
Understanding the Network Health Detail Report
Understanding the Network Health Report
Understanding the Node List Report
Understanding the Proxy Report
Understanding the Ram Journal Report
Understanding the Service Report

6.1 Understanding the Cache Size Report

The cache size report can be executed either on demand or it can be added as part of the report batch and the Caches should have the <unit-calculator> subelement of <local-scheme> set to BINARY. The cache size file is a tab delimited file that is prefixed with the date and hour in YYYYMMDDHH format and post fixed with -cache-size.txt. For example 2009013101-cache-size.txt would be created on January 31, 2009 at 1:00 AM. Table 6-1 describes the content of the Cache Size report.

Table 6-1 Contents of the Cache Size Report

Column	Data Type	Description
Batch Counter	Long	A sequential counter to help integrate information between related files. This value does reset when the reporter restarts and is not consistent across members. However, it is helpful when trying to integrate files.
Cache Name	String	The name of the cache.
Cache Size	Double	The number of objects in the cache.
Memory Bytes	Double	The number of bytes consumed by the objects in the cache. This does not include indexes or over head.
MemoryMB	Double	The MB consumed by the objects in the cache. This does not include indexes or over head.
Avg Object Size	Double	The Average memory consumed by each object.

6.2 Understanding the Cache Usage Report

The cache usage report provides information about cache usage (gets, puts, evictions, and so on). The report is a tab-delimited file that is prefixed with the date and hour in YYYYMMDDHH format and post fixed with -cache-usage.txt. For example, 2010013113-cache-usage.txt would be created on January 31, 2010 1:00 pm. Table 6-2 describes the content of the cache utilization report.

Table 6-2 Contents of the Cache Usage Report

Column	Data Type	Description
Batch Counter	Long	A sequential counter to help integrate information between related files. This value resets when the Reporter restarts, and is not consistent across members. However, it is helpful when trying to integrate files.
Report Time	Date	The system time when the report is executed.
Service	String	The name of the cache service.
Cache Name	String	The name of the cache.
Tier	String	Value can be either `front` or `back`. Describes whether the cache resides in the front-tier (local cache) or back tier (remote cache).
Total Puts	Double	The total number of puts for the cache across the cluster since the last time the report was executed.
Total Puts Milliseconds	Double	The total number of milliseconds spent per put() invocation (PutsMillis) across the cluster since the last time the report was executed.
Total Gets	Double	The total number of gets for the cache across the cluster since the last time the report was executed.
Total Gets Milliseconds	Double	The total number of milliseconds spent per `get()` invocation (GetsMillis) across the cluster since the last time the report was executed.
Total Hits	Double	The total number of hits for the cache across the cluster since the last time the report was executed.
Total Hits Milliseconds	Double	The total number of milliseconds spent per `get()` invocation that is a hit (HitsMillis) across the cluster since the last time the report was executed.
Total Misses	Double	The total number of misses for the cache across the cluster since the last time the report was executed.
Total Misses Milliseconds	Double	The total number of milliseconds spent per `get()` invocation that is a miss (MissesMillis) across the cluster since the last time the report was executed.
Total Writes	Double	The total number of storage writes for the cache across the cluster since the last time the report was executed.
Total Writes Milliseconds	Double	The total number of milliseconds spent in storage write operations (WritesMillis) across the cluster since the last time the report was executed.
Total Reads	Double	The total number of reads from a cache store for the cache across the cluster since the last time the report executed.
Total Read Milliseconds	Double	The total number milliseconds on cache store reads for the cache across the cluster since the last time the report executed.
Total Failures	Long	The total number of storage failures for the cache across the cluster since the last time the report was executed.
Total Queue	Long	The sum of the queue link sizes across the cluster.
evictions	Long	The total number of evictions for the cache across the cluster since the last time the report was executed.
Cache Prunes	Long	The total number of prunes for the cache across the cluster since the last time the report was executed.
Cache Prunes Milliseconds	Long	The total number of milliseconds spent in the prune operation (PrunesMillis) across the cluster since the last time the report was executed.

6.3 Understanding the Flash Journal Report

The flash journal report displays statistics that are used to determine how well data is being stored to flash memory. The report is a tab-delimited file that is prefixed with the date and hour in YYYYMMDDHH format and post fixed with -flashjournal.txt. For example, 2010013113-flashjournal.txt would be created on January 31, 2010 1:00 pm. Table 6-3 describes the content of the cache utilization report.

Table 6-3 Contents of the Flash Journal Report

Column	Type	Description
Batch Counter	Long	A sequential counter to help integrate information between related files. This value does reset when the reporter restarts and is not consistent across members. However, it is helpful when trying to integrate files.
Report Time	Date	The system time when the report executed.
Node ID	Long	The member for the flash journal statistics.
FileCount	Integer	The number of journal files that are currently in use.
BinaryStoreCount	Integer	The number of active `JournalBinaryStore` objects that are using this journal.
TotalDataSize	Long	The amount of data, in bytes, that are currently stored for this journal.
TotalFileSize	Long	The total size of all journal files for this journal.
BacklogCount	Integer	The number of serialized values that have yet to be persisted.
BacklogSize	Integer	The maximum size, in bytes, of the backlog. The backlog is the amount of serialized values that has yet to be persisted. Client threads are blocked if this limit is exceeded and remain blocked until the backlog recedes below this limit
PoolSize	Integer	The total size, in bytes, of all available buffers in the pool.

6.4 Understanding the Management Report

The Management report contains refresh statistics that are used to determine if the management framework is providing a timely view of management data for all MBeans. The management file is a tab delimited file that is prefixed with the date and hour in YYYYMMDDHH format and post fixed with -Management.txt. For example 2009013113-Management.txt would be created on January 31, 2009 at 1:00 PM. Table 6-4 describes the content of the Management report.

Table 6-4 Contents of the Management Report

Column	Type	Description
Batch Counter	Long	A sequential counter to help integrate information between related files. This value does reset when the reporter restarts and is not consistent across members. However, it is helpful when trying to integrate files.
Report Time	Date	The system time when the report executed.
Refresh Policy	String	The refresh policy that is currently set, which is used to determine how data for remote models are being refreshed.
Refresh Time	Date	The timestamp when this model was last retrieved from a corresponding member. For local servers it is the local time.
Refresh Excess Count	Long	The number of times the MBean server predictively refreshed information and the information was not accessed.
Refresh Count	Long	The total number of snapshots retrieved since the statistics were last reset.
Refresh Prediction Count	Long	The number of times the MBean server used a predictive algorithm to refresh MBean information.
Refresh Timeout Count	Long	The number of times this management member has timed out while attempting to refresh remote MBean attributes.

6.5 Understanding the Memory Status Report

The Memory Status report must be run as part of a report batch. The values are helpful in understanding memory consumption on each member and across the grid. For data to be included, cluster members must be configured to publish platform MBean information. The memory status file is a tab delimited file that is prefixed with the date and hour in YYYYMMDDHH format and post fixed with -memory-status.txt. For example 2009013115-memory-status.txt would be created on January 31, 2009 at 3:00 PM. Table 6-5 describes the content of the Memory Status report.

Table 6-5 Contents of the Memory Status Report

Column	Data Type	Description
Batch Counter	Long	A sequential counter to help integrate information between related files. This value does reset when the reporter restarts and is not consistent across members. However, it is helpful when trying to integrate files.
Report Time	Date	The system time when the report executed.
JVM Uptime	Long	The amount of time since the JVM was started.
Node Id	Long	The member for the memory statistics.
Gc Name	String	The name of the Garbage Collector information.
CollectionCount	Long	The number of garbage collections that have happened since the computer started.
Delta Collection Count	Long	The number of garbage collections that have occurred since the last execution of the report.
CollectTime	Long	The number of milliseconds the JVM has spent on garbage collection since the start of the JVM.
Delta Collect Time	Long	The number of milliseconds the JVM has spent on garbage collection since the last execution of the report.
Last GC Start Time	Long	The start time of the last garbage collection.
Last GC Duration Millis	Long	The total amount of time of the last garbage collection.
Last GC Stop Time	Long	The stop time of the last garbage collection.
Heap Committed	Long	The number of heap bytes committed at the time of report.
Heap Init	Long	The number of heap bytes initialized at the time of the report.
Heap Max	Long	The Maximum number of bytes used by the JVM since the start of the JVM.
Heap Used	Long	The bytes used by the JVM at the time of the report.

6.6 Understanding the Network Health Detail Report

The Network Health report supporting member-level details for determining the health of the network communications. The network health detail file is a tab delimited file that is prefixed with the date and hour in YYYYMMDDHH format and post fixed with -network-health-detail.txt. For example 2009013114-network-health.txt would be created on January 31, 2009 at 2:00PM. Table 6-6 describes the content of the Network Health Detail report.

Table 6-6 Contents of the Network Health Detail Report

Column	Data Type	Description
Batch Counter	Long	A sequential counter to help integrate information between related files. This value does reset when the reporter restarts and is not consistent across members. However, it is helpful when trying to integrate files.
Report Time	Date	The system time when the report executed.
RefreshTime	Date	The system time when management information was last retrieved from a corresponding node. Local servers display the local time.
Node Id	Long	The member for the network statistics.
Tx Success	Double	The publisher success rate for the member. If this value is within 2%-3% of the "Min Node Tx Success" and more than 10% less than the "Grid Tx Success" for the batch in the Network Health File, the corresponding member may be having difficulty communicating with the cluster. Constrained CPU, constrained network bandwidth or high network latency could cause this to occur.
RX Success	Double	The receiver success rate for the member. If this value is within 2%-3% of the "Min Node Rx Success" and more than 10% less than the "Grid Tx Success" for the batch in the Network Health File, the corresponding member may be having difficulty communicating with the cluster. Constrained CPU, constrained network bandwidth or high network latency could cause this to occur.
Packets Sent	Double	The total number of network packets sent by the member.
Current Packets Sent	Long	The number of packets sent by the member since the prior execution of the report.
Packets Resent	Long	The total number of network packets resent by the member. Packets are resent when the receiver of the packet receives and invalid packet or when an acknowledge packet is not sent within the appropriate amount of time.
Current Packet Resent	Long	The number of network packets resent by the member since the prior execution of the report.
PacketsRepeated	Long	The total number of packets received more than once.
Current Packets Repeated	Long	The number of packets received more than once since the last execution of the report.
Packets Received	Long	The total number of packets received by the member.
Current Packets Received	Long	The total number of packets received by the member since the last execution of the report.

6.7 Understanding the Network Health Report

The Network Health report contains the primary aggregates for determining the health of the network communications. The network health file is a tab delimited file that is prefixed with the date and hour in YYYYMMDDHH format and post fixed with -network-health.txt. For example 2009013113-network-health.txt would be created on January 31, 2009 at 1:00 PM. Table 6-7 describes the content of the Network Health report.

Table 6-7 Contents of the Network Health Report

Column	Type	Description
Batch Counter	Long	A sequential counter to help integrate information between related files. This value does reset when the reporter restarts and is not consistent across members. However, it is helpful when trying to integrate files.
Report Time	Date	The system time when the report executed.
Min Node Rx Success	Double	The minimum receiver success rate for a member in the cluster. If this value is considerably less (10%) than the Grid Rx Success rate. Further analysis using the Network Health Detail should be done.
Grid Rx Success	Double	The receiver success rate for the grid as a whole. If this value is below 90%. Further analysis of the network health detail should be done.
Min Node Tx Success	Double	The minimum publisher success rate for a member in the cluster. If this value is considerably less (10%) than the Grid Rx Success rate. Further analysis using the Network Health Detail should be done.
Grid TX Success	Double	The publisher success rate for the grid as a whole. If this value is below 90%. Further analysis of the network health detail should be done.

6.8 Understanding the Node List Report

Due to the transient nature of the node identifier (nodeId), the reporter logs out a list of members and user-defined member identity information. See "member-identity" in the Oracle Coherence Developer's Guide. The node list file is a tab delimited file that is prefixed with the date and hour in YYYYMMDDHH format and post fixed with -nodes.txt. For example 2009013101-nodes.txt would be created on January 31, 2009 at 1:00 AM. Table 6-8 describes the content of the Node List report.

Table 6-8 Contents of the Node List Report

Column	Data Type	Description
Batch Counter	Long	A sequential counter to help integrate information between related files. This value does reset when the reporter restarts and is not consistent across members. However, it is helpful when trying to integrate files.
Report Time	Date	The system time when the report executed.
Node Id	String	The numeric member identifier.
Unicast Address	String	The Unicast address for the member.
Member Name	String	The member name.
Process Name	String	The process name for the member.
Role Name	String	The role name for the member.
Machine Name	String	The computer name for the member.
Rack Name	String	The rack name for the member.
Site Name	String	The site name for the member.
Refresh Time	Date/Time	The time which the information was refreshed from a remote member. If the time is different than the refresh time on other rows in the batch, the member did not respond in a timely matter. This is often caused by a member preforming a garbage collection. Any information regarding a member with an "old" refresh date is questionable.

6.9 Understanding the Proxy Report

The proxy file provides information about proxy servers and the information being transferred to clients. The Proxy file is a tab delimited file that is prefixed with the date and hour in YYYYMMDDHH format and post fixed with -report-proxy.txt. For example 2009013101-report-proxy.txt would be created on January 31, 2009 at 1:00 AM. Table 6-9 describes the content of the Proxy report.

Table 6-9 Contents of the Proxy Report

Column	Type	Description
Batch Counter	Long	A sequential counter to help integrate information between related files. This value does reset when the reporter restarts and is not consistent across members. However, it is helpful when trying to integrate files.
Report Time	Date	The system time when the report executed.
Node Id	String	The numeric member identifier.
Service Name	String	The name of the proxy service.
HostIp	String	The IP Address and Port of the proxy service.
Connection Count	Long	The current number of connections to the proxy service.
Outgoing Byte Backlog	Long	The number of bytes queued to be sent by the proxy service.
Outgoing Message Backlog	Long	The number of messages queued by the proxy service.
Bytes Sent	Long	The number of bytes sent by the proxy service since the last execution of the report.
Bytes Received	Long	The number of bytes received by the proxy service since the last execution of the report.
Messages Sent	Long	The number of messages sent by the proxy service since the last execution of the report.
Messages Received	Long	The number of messages received by the proxy service since the last execution of the report.

6.10 Understanding the Ram Journal Report

The ram journal report displays statistics that are used to determine how well data is being stored to RAM memory. The report is a tab-delimited file that is prefixed with the date and hour in YYYYMMDDHH format and post fixed with -ramjournal.txt. For example, 2010013113-ramjournal.txt would be created on January 31, 2010 1:00 pm. Table 6-10 describes the content of the cache utilization report.

Table 6-10 Contents of the Ram Journal Report

Column	Type	Description
Batch Counter	Long	A sequential counter to help integrate information between related files. This value does reset when the reporter restarts and is not consistent across members. However, it is helpful when trying to integrate files.
Report Time	Date	The system time when the report executed.
Node ID	Long	The member for the flash journal statistics.
FileCount	Integer	The number of journal files that are currently in use.
BinaryStoreCount	Integer	The number of active `JournalBinaryStore` objects that are using this journal.
TotalDataSize	Long	The amount of data, in bytes, that are currently stored for this journal.
TotalFileSize	Long	The total size of all journal files for this journal.

6.11 Understanding the Service Report

The service report provides information to the requests processed, request failures, and request backlog, tasks processed, task failures and task backlog. Request Count and Task Count are useful to determine performance and throughput of the service. RequestPendingCount and Task Backlog are useful in determining capacity issues or blocked processes. Task Hung Count, Task Timeout Count, Thread Abandoned Count, Request Timeout Count are the number of unsuccessful executions that have occurred in the system. Table 6-11 describes the contents of the Service report.

Table 6-11 Contents of the Service Report

Column	Data Type	Description
Batch Counter	Long	A sequential counter to help integrate information between related files. This value does reset when the reporter restarts and is not consistent across members. However, it is helpful when trying to integrate files.
Report Time	Date	The system time when the report executed.
Service	String	The service name.
Node Id	String	The numeric member identifier.
Refresh Time	Date	The system time when the service information was updated from a remote member.
Request Count	Long	The number of requests since the last report execution.
RequestPendingCount	Long	The number of pending requests at the time of the report.
RequestPendingDuration	Long	The duration for the pending requests at the time of the report.
Request Timeout Count	Long	The number of request timeouts since the last report execution.
Task Count	Long	The number of tasks executed since the last report execution.
Task Backlog	Long	The task backlog at the time of the report execution.
Task Timeout Count	Long	The number of task timeouts since the last report execution.
Task Hung Count	Long	The number of tasks that hung since the last report execution.
Thread Abandoned Count	Long	The number of threads abandoned since the last report execution.