Telemetry Reporter Metrics

The following table shows the metadata collected by Confluent Telemetry Reporter and sent to Confluent to power Proactive Support.

Metric name io.confluent.kafka.server/.* Description
confluent_audit/audit_log_fallback_rate_per_minute Number of audit log fallbacks per minute.
confluent_audit/audit_log_rate_per_minute Number of audit logs created per minute.
confluent_authorizer/authorization_request_rate_per_minute Number of authorization requests per minute.
confluent_authorizer/authorization_allowed_rate_per_minute Number of authorizations allowed per minute.
confluent_authorizer/authorization_denied_rate_per_minute Number of authorizations denied per minute.
confluent_auth_store/rbac_role_bindings_count Number of RBAC role bindings.
confluent_auth_store/rbac_access_rules_count Number of RBAC access rules.
confluent_auth_store/acl_access_rules_count Number of ACL access rules.
acl_authorizer/zookeeper_disconnects/total/delta Kafka Authorizer
acl_authorizer/zookeeper_expires/total/delta  
broker_failure/zookeeper_disconnects/total/delta ZooKeeper client disconnect rate.
broker_failure/zookeeper_expires/total/delta ZooKeeper client expiration rate.
broker_topic/bytes_in/total/delta Topic bytes in rate.
broker_topic/bytes_out/total/delta Topic bytes our rate.
broker_topic/failed_produce_requests/total/delta Topic failed produce rate.
broker_topic/failed_fetch_requests/total/delta Topic failed fetch rate.
broker_topic/produce_message_conversions/total/delta Number of producer message conversions per second. Conversions occur when the broker and client message version don’t match.
broker_topic/fetch_message_conversions/total/delta Number of consumer message conversions per second. Conversions occur when the broker and client message version don’t match.
controller/active_controller_count Number of active controllers.
controller/leader_election_rate_and_time_ms Leader election rate and latency.
controller/offline_partitions_count Number of partitions that don’t have an active leader and aren’t writable or readable. Alert if value is greater than 0.
controller/unclean_leader_elections/total Total number of unclean leader elections.
controller_channel/connection_close_rate Number of connections closed per second in the window.
controller_channel/connection_close_total Total number of connections closed in the window.
controller_channel/connection_count Number of currently open connections to the broker.
controller_channel/connection_creation_rate Number of new connections established per second in the window.
controller_channel/connection_creation_total Total number of connections created in the window.
controller_channel/request_size_avg Average size of all requests in the window for a broker.
controller_channel/request_size_max Maximum size of any request sent in the window for a broker.
controller_channel_manager/queue_size  
controller_channel_manager/total_queue_size  
controller_event_manager/event_queue_size  
delayed_operation_purgatory/purgatory_size Number of requests waiting in the producer purgatory. Should be non-zero when acks=all is used on the producer.
executor/zookeeper_disconnects/total/delta ZooKeeper SBC client disconnect rate.
executor/zookeeper_expires/total/delta ZooKeeper SBC client disconnect rate.
fetch/queue_size  
group_coordinator/partition_load_time_max  
log_cleaner_manager/achieved_cleaning_ratio/time/delta Dirty ratio of topic partitions to be compacted, in percentage.
log_cleaner_manager/achieved_cleaning_ratio/total/delta  
log_cleaner_manager/compacted_partition_bytes Amount of data in compacted topics for each log directory.
log_cleaner_manager/max_dirty_percent The cleanable ratio of the dirtiest log.
log_cleaner_manager/time_since_last_run_ms Time since the last log cleaner run, in milliseconds. Resets to 0 when the log cleaner thread runs. Increases constantly if the log cleaner thread is not alive.
log_cleaner_manager/uncleanable_bytes Number of uncleanable bytes from uncleanable partitions for each log directory.
log_cleaner_manager/uncleanable_partitions_count Number of partitions marked as uncleanable for each log directory.
replica_alter_log_dirs_manager/max_lag  
replica_fetcher/request_size_avg Average size of requests to the Replica Manager.
replica_fetcher/request_size_max Size of the largest request to the Replica Manager.
replica_fetcher_manager/max_lag Maximum lag in messages between the follower and leader replicas. Controlled by the replica.lag.max.messages config.
replica_manager/isr_shrinks Number of ISR shrinks per second.
replica_manager/leader_count Number of leaders on the current broker. Usually, this value should be even, across all brokers. If it’s not, set auto.leader.rebalance.enable to true on all brokers in the cluster.
replica_manager/partition_count Number of partitions across all topics in the cluster.
replica_manager/under_min_isr_partition_count Number of partitions that have an in-sync replicas count less than minIsr. These partitions are not available to producers that use acks=all.
replica_manager/under_replicated_partitions Number of under-replicated partitions.
request/errors/total/delta  
request/local_time_ms/time/delta Time the request is processed at the leader, in milliseconds.
request/local_time_ms/total/delta  
request/queue_size  
request/remote_time_ms/time/delta Time the request waits for the follower. Non-zero for produce requests when acks=all.
request/remote_time_ms/total/delta  
request/request_queue_time_ms/time/delta Time that the request waits in the request queue, in milliseconds.
request/request_queue_time_ms/total/delta  
request/requests The average number of requests sent per second.
request/response_queue_time_ms/time/delta Time that the request waits in the response queue, in milliseconds.
request/response_queue_time_ms/total/delta  
request/response_send_time_ms/time/delta Time to send the response, in milliseconds.
request/response_send_time_ms/total/delta  
request/total_time_ms/time/delta Total time to serve the specified request, in milliseconds.
request/total_time_ms/total/delta  
request_channel/request_queue_size Size of the request queue. A congested request queue can’t process incoming or outgoing requests.
request_channel/response_queue_size Size of the response queue. The response queue is unbounded. A congested response queue can result in delayed response times and memory pressure on the broker.
request_handler_pool/request_handler_avg_idle_percent Average fraction of time the request handler threads are idle. Values are between 0 (all resources are used) and 1 (all resources are available).
session_expire_listener/zookeeper_disconnects/total/delta Number of disconnects per second.
session_expire_listener/zookeeper_expires/total/delta Number of session expirations per second.
socket_server/connections Number of currently open connections to the broker.
socket_server/successful_authentication_total/delta  
socket_server/failed_authentication_total/delta  
socket_server/network_processor_avg_idle_percent Average fraction of time that the network processor threads are idle. Values are between 0 (all resources are used) and 1 (all resources are available).
socket_server/request_size_avg Average size of requests to the broker.
socket_server/request_size_max Size of the largest request to the broker.

Metadata tags

Additional metadata tags are attached to each metric record received, enables storing and associating the metric with the correct resource.

Tag name Description
kafka.cluster.id Unique Kafka cluster ID
kafka.version Kafka version of broker
kafka.broker.id Unique broker identifier in Kafka cluster
java.version Java version of broker
host.hostname Host name for broker
topic Topic name. Populated when collecting topic-level metrics