eG Monitoring
 

Measures reported by KafkaProdInsTest

A Kafka client publishes records to the Kafka cluster (broker). The producer is thread safe and it shares a single producer instance across threads that would generally be faster than having multiple instances. This test monitors those producer instances and reports potential contentions, if any.

Outputs of the test: One set of results for the target Apache Kafka Message System.

The measures made by this test are as follows:

Measurement Description Measurement Unit Interpretation
batchSizeAvg Indicates the average number of bytes sent per partition per-request. MB

 

batchSplitRate Indicates the average number of batch splits. Splits/sec

 

bufferAvailableBytes Indicates the total amount of buffer memory that is not being used (either unallocated or in the free list). MB

 

bufferExhaustedRate Indicates the average number of sent record, that are dropped due to buffer exhaustion. Number

A very low is required for this measure.

bufferTotalBytes Indicates the maximum amount of buffer memory the client can use (whether or not it is currently used). MB

 

bufferpoolWaitRatio Indicates the fraction of time an appender waits for space allocation.

Number

 

compressionRateAvg Indicates the average compression rate of record batches and it is defined as the average ratio of the compressed batch size over the uncompressed size.

Number/sec

 

connectionCloseRate Indicates the closed connections.

Connections/sec

 

connectionCount Indicates the current number of active connections.

Number

A very high value is required for this measure.

connectionCreationRate Indicates the new established connections.

Connections/sec

 

failedAuthenticationRate Indicates the connections for failed authentication.

Connections/sec

Ideally, the value of these measures should be zero.

failedReauthentication Indicates the connections for failed re-authentication.

Connections/sec

Ideally, the value of these measures should be zero.

ioTimeNsAvg Indicates the average length of time for I/O per select call.

Seconds

 

ioWaitTimeNsAvg Indicates the average length of time the I/O thread spent waiting for a socket ready for reads or writes.

Seconds

 

metadataAge Indicates the age of the current producer metadata being used.

Seconds

 

waitingThreads ndicates the number of user threads blocked waiting for buffer memory to enqueue their records.

Number