Kafka Broker Details

The Broker Details page displays detailed information of a selected broker on a node. This page consists of two tabs: Broker, and Node.

Broker Tab

The Broker tab displays charts related to the broker service running on the node.

ChartDescription
Avg CPU UtilizationThe average amount of CPU utilized (in %) by a broker on a host.
Avg Disk WaitThe Avg Disk Wait graph displays the average amount of time (in %) the disk waited to become available to a broker to perform a read/write operation. The graph also displays the Read Time and Write Time (in ms).
Disk UtilizationThe amount of disk utilized and the total allocated disk space on the host.

Request Metrics Chart

The Request Metrics chart displays the following metrics of incoming and outgoing data per second.

  • Requests per Second: The number of incoming and outgoing requests per second.
  • Time Taken: The time taken to process the incoming and outgoing requests.

You can view the 95th percentile of the time range you selected and the overall 95th percentile of the following metrics in the chart.

  • Request Queue Time: The time elapsed waiting in the request queue.
  • Local Time: The time spent to process the request in the same node.
  • Remote Time: The time spent to process the request by a broker in another node.
  • Throttle Time Response Queue Time: The amount of time (in ms) the client was throttled in queue.
  • Response Send Time: The time taken to send the response.

Top Topics Chart

The Top Topics chart displays the highest number of topics by the following filters.

note

You can sort by the following filters by clicking the drop-down in the top right of the Top Topics graph and selecting a filter.

MetricDescription
Bytes in per topicThe size of incoming messages in each topic.
Bytes out per topicThe size of outgoing messages from each topic.
Messages in per topicThe number of incoming messages to each topic.
Produce Requests per topicThe number of producer requests for each topic.

Viewing Metrics of a Topic

To view the Partition Size and Consumer Lag for each topic, click the topic in the first chart of the Top Topics chart.

JVM Usage Chart

The JVM Usage chart displays the amount of JVM memory used by a broker based on the following metrics.

  • Max Heap: The maximum amount of heap memory allocated by the JVM.
  • Committed Non Heap: The amount of allocated memory other than the heap memory in the JVM.
  • Heap Used: The amount of heap memory used by the JVM.
  • Non Heap Memory: The amount of memory used other than the heap memory in the JVM.

JVM GC Chart

The JVM GC chart displays the amount of JVM used in garbage collection while executing a broker task for a number of times for a certain duration. You can monitor the following metrics in the JVM GC chart.

  • Young Generation GC Time: The time taken by GC to collect new objects.
  • Young Generation GC Count: The number of new objects in the GC.
  • Old Generation GC Time: The time taken by the GC to collect longer surviving objects.
  • Old Generation GC Count: The number of longer surviving objects in the GC.

Broker Load Chart

The Broker Load chart displays the load balancing of broker across the following metrics.

-System Load: The system load value can be between 0.0 and 1.0. A value of 0.0 indicates that all CPUs were idle recently, while a value of 1.0 means indicates that all CPUs were actively running 100% of the time recently. -Broker Load: The broker load value can be between 0.0 and 1.0. A value of 0.0 indicates that none of the CPUs were running any threads from the JVM process recently. A value of 1.0 indicates that all CPUs were actively running application threads and internal threads from the JVM 100% of the time recently.

Node Tab

The Node tab displays memory-related and network-related statistics of a node that the broker runs on.

Summary Panel

The Summary Panel displays the following metrics.

  • Up Time: The time that the node is available and working.
  • CPUs: The number of CPUs in the node.
  • Total Memory: The total memory allocated to the node.
  • Memory Used %: The amount of memory used (in %).
  • CPU Used %: The amount of CPU used (in %).
  • Disk Used %: The amount of disk used (in %).

Charts

The following charts are displayed for a node.

ChartDescription
CPU UsageThe amount of CPU utilized by the node.
Network Usage Per SecThe network throughput measured for every second.
Memory UsageThe amount of memory utilized by the node.

IO Usage Charts

The IO Usage charts describe the number of input bytes read and the number of output bytes written to a node. The following IO metrics are displayed.

note

Click Show Per Mount Usage to view I/O usage on every mount in the node.

IO MetricDescription
I/O OpsThe input/output operations per second. This chart displays the data for Read iops and Write iops.
I/O TimeThe time utilized by input/output operations. This chart displays Read Time and Write Time (in ms) of I/O operations in a node.
I/O BytesThe size of I/O operations. This chart displays the size of incoming and outgoing operations in the form of Read Bytes and Write Bytes.
I/O Request TimeThe time taken by I/O requests waiting in queue to execute.