This guide provides a detailed overview of the Topic Monitoring view within Kadeck's Apache Kafka UI, allowing users to monitor and analyze various metrics associated with Kafka topics.
Overview
Topic select: Dropdown to select specific topics or view metrics for all topics collectively.
Topics: The total number of topics.
Partitions: The total number of partitions across all topics.
Messages: The cumulative count of messages across topics.
U/RP: Under-replicated partitions count.
Partition skew: A measure indicating the imbalance in data distribution across partitions.
Produced and Consumed Metrics
Produced per topic (msg/s): Graph showcasing the message production rate over time.
Total produced: Total messages produced during the specified timeframe.
Latest (selection): Most recent data point for message production.
Consumed per topic (msg/s): Graph depicting the message consumption rate over time.
Total read: Total messages consumed during the selected period.
Latest (selection): Most recent data point for message consumption.
Topic Size
This section provides a visual representation of the size (in messages) of various topics. Larger blocks indicate topics with a higher volume of messages.
Partition Distribution
Partition distribution (msg): Graphical representation of the distribution of messages across various partitions.
Partition skew: A line graph showing the skewness of partition data distribution over time. High skewness indicates imbalances in message distribution among partitions.
Latest (selection): Most recent skewness value.
Partition distribution (leader disk): Visual representation of the distribution of partitions based on the leader disk.
Tips for Optimal Usage
Consistent Monitoring: Regularly monitor the Produced and Consumed metrics to ensure balanced production and consumption rates.
Address Skewness: High partition skew can lead to bottlenecks. Address any noticeable skew to maintain efficient data processing.
Topic Size: Keep an eye on growing topics to ensure that they don't overwhelm the system or lead to increased latency.