You can view broker metrics by selecting the Broker on the left side of the screen . This is a screen that allows you to see the current key metrics of all brokers at a glance and is composed of several detailed items.
Overall cluster health
- Number of online brokers. Shows the number of operating brokers
- Active Controllers. Displays the number of active controllers among brokers. Error if this value is not 1.
- Number of topics. Cluster-wide topic count.
- Number of active partitions. Number of active partitions.
- Zookeeper connection status. Displays the online broker's ZooKeeper connection status. If the number of online brokers does not match the number of connections, an error is displayed.
- Under ISR. Number of topic partitions that do not meet the min ISR setting. May temporarily be non-zero.
- Under replicated partition. Number of topic partitions that do not meet the replication factor. A value of 0 is normal.
Check under replicated partitions
Select the Under replicated partition warning link to see a list of topic partitions that do not meet the current replication factor.
Check Under ISR Partitions
Select the Under ISR warning link to see a list of topic partitions that do not meet the current min ISR.
Message creation/consumption
- PRODUCING. Cluster-wide message generation bytes per second.
- CONSUMING. Cluster-wide number of message bytes consumed per second. Includes Consumer and Follower.
Node resource average utilization
- CPU average utilization. Broker-wide CPU utilization average.
- Memory average utilization. Broker-wide memory utilization average.
- Network average utilization. Broker-wide network utilization average.
- Request handler average utilization. Broker-wide request handler average utilization.
Topic Disk Usage
- Total disk usage. Total cluster topic partition usage total.
- Max usage. Maximum of total topic partition usage by broker.
- Min usage. Minimum of the total topic partition usage for each broker.
- Distribution map. If the topic partition usage for each broker differs by more than 10%, it is considered unbalanced. If the disk usage is unbalanced, the Smart Rebalance button is displayed. Please refer to the Smart Rebalance chapter.
Only the amount of disk used by the Kafka broker to store topic partitions is counted.
Broker List
A list of all brokers is displayed at the bottom of the broker metrics overview page. You can view the metrics details screen for a broker by clicking the arrow for its host name in the list.
- Broker name. Broker host name, IP:PORT information. Separate label for controllers.
- Status. Displays the running status of the broker.
- Partitions. Number of partitions stored in that broker.
- Throughput. The number of message bytes produced/consumed per second for that broker.
- CPU Load. CPU utilization of that broker.
- Node disk usage. Average disk usage for that broker. Average value of disk usage items by mount.
- Disk usage by mount (accessible through View Details). Disk usage by major mount points for that broker.