Pulsar metrics

Pulsar exposes the following metrics in Prometheus format. You can monitor your clusters with those metrics.

• • Pulsar Functions
  • Connectors
  • Proxy
  • Pulsar transaction

The following types of metrics are available:

• Counter: a cumulative metric that represents a single monotonically increasing counter. The value increases by default. You can reset the value to zero or restart your cluster.
• Gauge: a metric that represents a single numerical value that can arbitrarily go up and down.
• Histogram: a histogram samples observations (usually things like request durations or response sizes) and counts them in configurable buckets. The _bucket suffix is the number of observations within a histogram bucket, configured with parameter {le="<upper inclusive bound>"}. The _count suffix is the number of observations, shown as a time series and behaves like a counter. The _sum suffix is the sum of observed values, also shown as a time series and behaves like a counter. These suffixes are together denoted by _* in this doc.
• Summary: similar to a histogram, a summary samples observations (usually things like request durations and response sizes). While it also provides a total count of observations and a sum of all observed values, it calculates configurable quantiles over a sliding time window.

This reference lists Pulsar metrics and selected metrics from its dependencies. Enabled components and configuration determine which series are present. Prometheus counters registered through the Java client expose a _total sample suffix, even when a collector's base name omits it. Summaries expose quantiles plus _count and _sum; classic histograms expose _bucket, _count, and _sum. The explicitly named Pulsar *_le_* bucket gauges are separate metric families, not classic Prometheus histograms. See metrics changes when upgrading.

Matching the scrape interval to the stats periods

Match the Prometheus scrape interval to the stats periods

Many metrics that brokers and bookies expose to Prometheus are not cumulative. The process computes them over a fixed period, 60 seconds by default, and then starts over. These include topic and namespace rates, latency summaries and buckets, and managed-ledger period statistics.

Cumulative counters can be scraped at any interval. For per-period metrics, match the Prometheus scrape interval to the stats period:

• A scrape interval shorter than the period reads the same values several times. The metrics change only once per period, however often you scrape.
• A scrape interval longer than the period misses periods. Latency distributions and rates then cover only some of the time.

For finer-grained metrics, such as a 5-second or 10-second scrape during load tests, set the following periods to the scrape interval. All values below are in seconds.

Broker (conf/broker.conf):

Setting	Default	Purpose
statsUpdateFrequencyInSecs	60	Updates topic and namespace rates and latency summaries for each period.
statsUpdateInitialDelayInSecs	60	Delays the first stats update after startup; set it to the same value as the update period.
managedLedgerStatsPeriodSeconds	60	Sets the period for managed-ledger statistics (pulsar_ml_*), including latency and entry-size buckets.
managedLedgerPrometheusStatsLatencyRolloverSeconds	60	Sets the rollover period for the BookKeeper client's latency statistics, exposed when bookkeeperClientExposeStatsToPrometheus=true.

Bookie (conf/bookkeeper.conf):

Setting	Default	Purpose
prometheusStatsLatencyRolloverSeconds	60	Sets the rollover period for the bookie's request, journal, and storage latency statistics.

The default bookkeeper.conf does not list prometheusStatsLatencyRolloverSeconds, so add it to the file. With the Pulsar Docker image, pass it as the environment variable PULSAR_PREFIX_prometheusStatsLatencyRolloverSeconds; the PULSAR_PREFIX_ form adds keys that are absent from the file. With the Pulsar Helm chart, set this environment variable through bookie.configData, as described in Component configuration.

For example, configure a 10-second scrape interval as follows:

# conf/broker.conf
statsUpdateFrequencyInSecs=10
statsUpdateInitialDelayInSecs=10
managedLedgerStatsPeriodSeconds=10
managedLedgerPrometheusStatsLatencyRolloverSeconds=10

# conf/bookkeeper.conf
prometheusStatsLatencyRolloverSeconds=10

# prometheus.yml
scrape_configs:
  - job_name: pulsar
    scrape_interval: 10s
    static_configs:
      - targets: ["broker:8080", "bookie:8000"]

Replace the example targets with your broker and bookie metrics endpoints, or use your existing Prometheus service discovery configuration.

Shorter periods cost more broker CPU because each stats update goes through every topic. On brokers with many topics, keep the default 60-second period for production monitoring and use shorter periods for load tests and troubleshooting.

ZooKeeper

The ZooKeeper metrics are exposed under "/metrics" at port 8000. You can use a different port by configuring the metricsProvider.httpPort in conf/zookeeper.conf.

ZooKeeper provides a New Metrics System since 3.6.0. For more detailed metrics, refer to the ZooKeeper Monitor Guide.

BookKeeper

The BookKeeper metrics are exposed under "/metrics" at port 8000. You can change the port by updating prometheusStatsHttpPort in the bookkeeper.conf configuration file.

Server metrics

Name	Type	Description
bookie_SERVER_STATUS	Gauge	The server status for bookie server. 1: the bookie is running in writable mode. 0: the bookie is running in read-only mode.
bookkeeper_server_ADD_ENTRY_count	Counter	The total number of ADD_ENTRY requests received at the bookie. The success label is used to distinguish between successes and failures.
bookkeeper_server_READ_ENTRY_count	Counter	The total number of READ_ENTRY requests received at the bookie. The success label is used to distinguish between successes and failures.
bookie_WRITE_BYTES	Counter	The total number of bytes written to the bookie.
bookie_READ_BYTES	Counter	The total number of bytes read from the bookie.
bookkeeper_server_ADD_ENTRY_REQUEST	Summary	The summary of request latency of ADD_ENTRY requests at the bookie. The success label is used to distinguish between successes and failures.
bookkeeper_server_READ_ENTRY_REQUEST	Summary	The summary of request latency of READ_ENTRY requests at the bookie. The success label is used to distinguish between successes and failures.
bookkeeper_server_BookieReadThreadPool_queue_{thread_id}	Gauge	The number of requests to be processed in a read thread queue.
bookkeeper_server_BookieReadThreadPool_task_queued	Summary	The waiting time of a task to be processed in a read thread queue.
bookkeeper_server_BookieReadThreadPool_task_execution	Summary	The execution time of a task in a read thread queue.

Journal metrics

Name	Type	Description
bookie_journal_JOURNAL_SYNC_count	Counter	The total number of journal fsync operations happening at the bookie. The success label is used to distinguish between successes and failures.
bookie_journal_JOURNAL_QUEUE_SIZE	Gauge	The total number of requests pending in the journal queue.
bookie_journal_JOURNAL_FORCE_WRITE_QUEUE_SIZE	Gauge	The total number of force write (fsync) requests pending in the force-write queue.
bookie_journal_JOURNAL_CB_QUEUE_SIZE	Gauge	The total number of callbacks pending in the callback queue.
bookie_journal_JOURNAL_ADD_ENTRY	Summary	The summary of request latency of adding entries to the journal.
bookie_journal_JOURNAL_SYNC	Summary	The summary of fsync latency of syncing data to the journal disk.
bookie_journal_JOURNAL_CREATION_LATENCY	Summary	The latency created by a journal log file.

Storage metrics

Name	Type	Description
bookie_ledgers_count	Gauge	The total number of ledgers stored in the bookie.
bookie_entries_count	Gauge	The total number of entries stored in the bookie.
bookie_write_cache_size	Gauge	The bookie write cache size (in bytes).
bookie_read_cache_size	Gauge	The bookie read cache size (in bytes).
bookie_DELETED_LEDGER_TOTAL	Counter	The total number of ledgers deleted since the bookie has started.
bookie_ledger_writable_dirs	Gauge	The number of writable directories in the bookie.
bookie_flush	Gauge	The table flush latency of bookie memory.
bookie_throttled_write_requests	Counter	The number of write requests to be throttled.

Replication metrics

Name	Type	Description
auditor_NUM_UNDER_REPLICATED_LEDGERS	Summary	The distribution of num under_replicated ledgers on each auditor run.
auditor_UNDER_REPLICATED_LEDGERS_TOTAL_SIZE	Summary	The distribution of under_replicated ledgers total size on each auditor run.
auditor_URL_PUBLISH_TIME_FOR_LOST_BOOKIE	Summary	The latency distribution of publishing under replicated ledgers for lost bookies.
auditor_BOOKIE_TO_LEDGERS_MAP_CREATION_TIME	Summary	The latency distribution of creating bookies-to-ledgers map.
auditor_CHECK_ALL_LEDGERS_TIME	Summary	The latency distribution of checking all ledgers.
auditor_PLACEMENT_POLICY_CHECK_TIME	Summary	The latency distribution of placementPolicy check.
auditor_REPLICAS_CHECK_TIME	Summary	The latency distribution of replicas check.
auditor_AUDIT_BOOKIES_TIME	Summary	The latency distribution of auditing all the bookies.
auditor_NUM_LEDGERS_CHECKED	Counter	The number of ledgers checked by the auditor.
auditor_NUM_FRAGMENTS_PER_LEDGER	Summary	The distribution of number of fragments per ledger.
auditor_NUM_BOOKIES_PER_LEDGER	Summary	The distribution of number of bookies per ledger.
auditor_NUM_BOOKIE_AUDITS_DELAYED	Counter	The number of bookie-audits delayed.
auditor_NUM_DELAYED_BOOKIE_AUDITS_DELAYES_CANCELLED	Counter	The number of delayed-bookie-audits cancelled.
auditor_NUM_LEDGERS_NOT_ADHERING_TO_PLACEMENT_POLICY	Gauge	The number of ledgers not adhering to placement policy found in placement policy check.
auditor_NUM_LEDGERS_SOFTLY_ADHERING_TO_PLACEMENT_POLICY	Gauge	The number of ledgers softly adhering to placement policy found in placement policy check.
auditor_NUM_UNDERREPLICATED_LEDGERS_ELAPSED_RECOVERY_GRACE_PERIOD	Gauge	The number of underreplicated ledgers elapsed recovery grace period.
auditor_NUM_LEDGERS_HAVING_NO_REPLICA_OF_AN_ENTRY	Gauge	The number of ledgers having an entry with all the replicas missing.
auditor_NUM_LEDGERS_HAVING_LESS_THAN_AQ_REPLICAS_OF_AN_ENTRY	Gauge	The number of ledgers having an entry with less than AQ number of replicas, this doesn't include ledgers counted towards numLedgersHavingNoReplicaOfAnEntry.
auditor_NUM_LEDGERS_HAVING_LESS_THAN_WQ_REPLICAS_OF_AN_ENTRY	Gauge	The number of ledgers having an entry with less than WQ number of replicas, this doesn't include ledgers counted towards numLedgersHavingLessThanAQReplicasOfAnEntry.
election_attempts	Counter	The number of auditor election attempts.
replication_worker_NUM_BYTES_READ	Summary	The distribution size of entries read by the replicator.
replication_worker_NUM_ENTRIES_READ	Counter	Number of entries read by the replicator.
replication_worker_NUM_ENTRIES_WRITTEN	Counter	Number of entries written by the replicator.
replication_worker_NUM_BYTES_WRITTEN	Summary	The distribution size of entries written by the replicator.
replication_worker_READ_DATA_LATENCY	Summary	The distribution of latency of read entries by the replicator.
replication_worker_WRITE_DATA_LATENCY	Summary	The distribution of latency of write entries by the replicator.
replication_worker_exceptions	Summary	Replication related exceptions.
replication_worker_rereplicate	Summary	Operation stats of re-replicating ledgers.
replication_worker_NUM_FULL_OR_PARTIAL_LEDGERS_REPLICATED	Counter	The number of ledgers re-replicated.
replication_worker_NUM_DEFER_LEDGER_LOCK_RELEASE_OF_FAILED_LEDGER	Counter	The number of defer-ledger-lock-releases of failed ledgers.
replication_worker_NUM_ENTRIES_UNABLE_TO_READ_FOR_REPLICATION	Counter	The number of entries ReplicationWorker unable to read.

Broker

The broker metrics are exposed under "/metrics/" at port 8080. You can change the port by updating webServicePort to a different port in the broker.conf configuration file.

All the metrics exposed by a broker are labeled with cluster=\$\{pulsar_cluster\}. The name of Pulsar cluster is the value of \$\{pulsar_cluster\}, which you have configured in the broker.conf file.

Metric exposure

exposeTopicLevelMetricsInPrometheus=true exports topic and subscription series; setting it to false exports namespace aggregates instead. This setting is dynamic:

bin/pulsar-admin brokers update-dynamic-config \
  --config exposeTopicLevelMetricsInPrometheus --value false

Change it back to true to restore topic-level series. Review dashboards and scrape cardinality when changing the aggregation level. To remove the dynamic override and use the broker configuration again, run pulsar-admin brokers delete-dynamic-config --config exposeTopicLevelMetricsInPrometheus.

For custom topic labels, see Configure topic metric labels.

Broker metrics

All the broker metrics are labeled with the following labels:

• cluster: cluster=${pulsar_cluster}. ${pulsar_cluster} is the cluster name that you have configured in the broker.conf file.

Name	Type	Description
pulsar_ml_cache_evictions	Gauge	The number of cache evictions during the last minute.
pulsar_ml_cache_inserted_entries_total	Counter	The number of entries inserted into the entry cache.
pulsar_ml_cache_evicted_entries_total	Counter	The number of entries evicted from the entry cache.
pulsar_ml_cache_entries	Gauge	The number of entries in the entry cache.
pulsar_ml_cache_hits_rate	Gauge	The number of cache hits per second on the broker side.
pulsar_ml_cache_hits_throughput	Gauge	The amount of data (byte per second) retrieved from the cache on the broker side.
pulsar_ml_cache_misses_rate	Gauge	The number of cache missed per second on the broker side.
pulsar_ml_cache_misses_throughput	Gauge	The amount of data (byte per second) that cannot be retrieved from the cache on the broker side.
pulsar_ml_cache_pool_active_allocations	Gauge	The number of currently active allocations in direct arena.
pulsar_ml_cache_pool_active_allocations_huge	Gauge	The number of currently active huge allocation in direct arena.
pulsar_ml_cache_pool_active_allocations_normal	Gauge	The number of currently active normal allocations in direct arena.
pulsar_ml_cache_pool_active_allocations_small	Gauge	The number of currently active small allocations in direct arena.
pulsar_ml_cache_pool_allocated	Gauge	Direct memory reserved by the managed-ledger cache allocator, in bytes.
pulsar_ml_cache_pool_used	Gauge	Direct memory in use by the managed-ledger cache allocator, in bytes.
pulsar_default_pool_allocated	Gauge	Direct memory reserved by the default Pulsar allocator, in bytes.
pulsar_default_pool_used	Gauge	Direct memory in use by the default Pulsar allocator, in bytes.
pulsar_ml_cache_used_size	Gauge	The size used to store the payloads of entries (in bytes).
pulsar_ml_count	Gauge	The number of currently opened managed ledgers.
topic_load_times (deprecated)	Summary	The topic load latency calculated in milliseconds.
pulsar_topic_load_times	Summary	The topic load latency calculated in milliseconds.
pulsar_broker_topics_count	Gauge	The number of Pulsar topics in this broker.
pulsar_broker_subscriptions_count	Gauge	The number of Pulsar subscriptions in this broker.
pulsar_broker_producers_count	Gauge	The number of active producers connected to this broker.
pulsar_broker_consumers_count	Gauge	The number of active consumers connected to this broker.
pulsar_broker_rate_in	Gauge	The total message rate coming into this broker (message per second).
pulsar_broker_rate_out	Gauge	The total message rate going out from this broker (message per second).
pulsar_broker_throughput_in	Gauge	The total throughput coming into this broker (byte per second).
pulsar_broker_throughput_out	Gauge	The total throughput going out from this broker (byte per second).
pulsar_broker_storage_size	Gauge	The total storage size of all topics in this broker (in bytes).
pulsar_broker_storage_logical_size	Gauge	The storage size of all topics in this broker without replicas (in bytes).
pulsar_broker_storage_write_rate	Gauge	The total message batches (entries) written to the storage for this broker (message batch per second).
pulsar_broker_storage_read_rate	Gauge	The total message batches (entries) read from the storage for this broker (message batch per second).
pulsar_broker_msg_backlog	Gauge	The total number of message backlogs in this broker (entries).
pulsar_storage_backlog_quota_check_duration_seconds	Histogram	The duration of the backlog quota check process (in seconds)
pulsar_broker_storage_backlog_quota_exceeded_evictions_total	Counter	The number of times a backlog was evicted since it has exceeded its quota. Includes the label quota_type = (time | size)

The pulsar_ml_cache_pool_* metrics describe the separate ml-cache allocator, while pulsar_default_pool_* describe the default allocator. The startup scripts select Netty's adaptive allocator for both. For adaptive and unpooled allocators, active allocation counts are unsupported and reported as -1; the allocated and used byte values both use the allocator's reported direct-memory usage. With the pooled allocator, allocation counts and the distinction between reserved and used pool bytes remain available. Compare these values with pulsar_ml_cache_used_size, which measures logical cached entry bytes: it does not include all memory retained by the allocator.

Additional broker metrics

These metrics are broker-wide. Publish latency is measured in milliseconds; its quantiles describe the stats rollover period, while its count and sum are cumulative.

Name	Type	Description
pulsar_broker_publish_latency	Summary	End-to-end broker publish latency, including broker-side completion. Quantiles: 0, 0.5, 0.95, 0.99, 0.999, 0.9999, and 1.
pulsar_broker_in_bytes_total	Gauge	Total bytes received across topics on the broker.
pulsar_broker_out_bytes_total	Gauge	Total bytes dispatched across topics on the broker.
pulsar_broker_storage_read_cache_misses_rate	Gauge	Storage read operations per second that miss the managed-ledger cache.
pulsar_broker_pending_bytes_to_dispatch	Gauge	Bytes loaded in memory and waiting to be dispatched to consumers.
pulsar_broker_http_rejected_requests_total	Counter	HTTP requests rejected by request rate limiting.
pulsar_health	Gauge	Result of the latest broker health check: 1 for success, 0 for failure, or -1 before a result is available.
pulsar_version_info	Gauge	Constant 1, labeled with the Pulsar version and commit.
pulsar_broker_load_manager_bundle_assigment	Summary	Time in milliseconds to select a broker for bundle assignment in the modular load manager. The exported name retains the spelling assigment.

BookKeeper client metrics

The BookKeeper client metrics is disabled by default. Set bookkeeperClientExposeStatsToPrometheus=true in broker.conf to expose the BookKeeper client metrics. Suppose you want to expose PerChannel BookKeeper client metrics, which allows you to get the metrics for each bookie that the BookKeeper client connected. You can set bookkeeperClientLimitStatsLogging=true in broker.conf. Note that enabling the PerChannel BookKeeper client metrics might introduce performance issues in high-load situations.

All the BookKeeper client metrics are labeled with the following label:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you configured in broker.conf.

Name	Type	Description
pulsar_managedLedger_client_bookkeeper_client_BOOKIE_QUARANTINE	Counter	The number of bookie clients to be quarantined.
pulsar_managedLedger_client_BookKeeperClientWorker_task_execution	Summary	The task execution latency calculated in milliseconds
pulsar_managedLedger_client_BookKeeperClientWorker_task_queued	Summary	The task queued latency calculated in milliseconds
pulsar_managedLedger_client_bookkeeper_client_ADD_ENTRY	Summary	Add entry latency calculated in milliseconds
pulsar_managedLedger_client_bookkeeper_client_READ_ENTRY	Summary	Read entry latency calculated in milliseconds
pulsar_managedLedger_client_bookkeeper_client_READ_LAC	Summary	Read Last Add Confirmed request latency calculated in milliseconds
pulsar_managedLedger_client_bookkeeper_client_WRITE_LAC	Summary	Write Last Add Confirmed request latency calculated in milliseconds

Namespace metrics

Namespace metrics are only exposed when exposeTopicLevelMetricsInPrometheus is set to false.

All the namespace metrics are labeled with the following labels:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you configured in broker.conf.
• namespace: namespace=\$\{pulsar_namespace\}. \$\{pulsar_namespace\} is the namespace name.

Name	Type	Description
pulsar_topics_count	Gauge	The number of Pulsar topics of the namespace owned by this broker.
pulsar_subscriptions_count	Gauge	The number of Pulsar subscriptions of the namespace served by this broker.
pulsar_producers_count	Gauge	The number of active producers of the namespace connected to this broker.
pulsar_consumers_count	Gauge	The number of active consumers of the namespace connected to this broker.
pulsar_rate_in	Gauge	The total message rate of the namespace coming into this broker (message per second).
pulsar_rate_out	Gauge	The total message rate of the namespace going out from this broker (message per second).
pulsar_throughput_in	Gauge	The total throughput of the namespace coming into this broker (byte per second).
pulsar_throughput_out	Gauge	The total throughput of the namespace going out from this broker (byte per second).
pulsar_consumer_msg_ack_rate	Gauge	The total message acknowledgment rate of the namespace owned by this broker (message per second).
pulsar_storage_size	Gauge	The total storage size of the topics in this namespace owned by this broker (bytes).
pulsar_storage_logical_size	Gauge	The storage size of topics in the namespace owned by the broker without replicas (in bytes).
pulsar_storage_backlog_size	Gauge	The total backlog size of the topics of this namespace owned by this broker (in bytes).
pulsar_storage_backlog_quota_exceeded_evictions_total	Counter	The number of times a backlog was evicted since it has exceeded its quota (When backlog quota's retention policy was set to consumer_backlog_eviction). Also includes label quota_type = (time | size)
pulsar_storage_offloaded_size	Gauge	The total amount of the data in this namespace offloaded to the tiered storage (bytes).
pulsar_storage_write_rate	Gauge	The total message batches (entries) written to the storage for this namespace (message batch per second).
pulsar_storage_read_rate	Gauge	The total message batches (entries) read from the storage for this namespace (message batch per second).
pulsar_subscription_delayed	Gauge	The total message batches (entries) are delayed for dispatching.
pulsar_storage_write_latency_le_*	Histogram	The entry rate of a namespace that the storage write latency is smaller with a given threshold. Available thresholds: pulsar_storage_write_latency_le_0_5: <= 0.5ms pulsar_storage_write_latency_le_1: <= 1ms pulsar_storage_write_latency_le_5: <= 5ms pulsar_storage_write_latency_le_10: <= 10ms pulsar_storage_write_latency_le_20: <= 20ms pulsar_storage_write_latency_le_50: <= 50ms pulsar_storage_write_latency_le_100: <= 100ms pulsar_storage_write_latency_le_200: <= 200ms pulsar_storage_write_latency_le_1000: <= 1s pulsar_storage_write_latency_le_overflow: > 1s
pulsar_entry_size_le_*	Histogram	The entry rate of a namespace that the entry size is smaller with a given threshold. Available thresholds: pulsar_entry_size_le_128: <= 128 bytes pulsar_entry_size_le_512: <= 512 bytes pulsar_entry_size_le_1_kb: <= 1 KB pulsar_entry_size_le_2_kb: <= 2 KB pulsar_entry_size_le_4_kb: <= 4 KB pulsar_entry_size_le_16_kb: <= 16 KB pulsar_entry_size_le_100_kb: <= 100 KB pulsar_entry_size_le_1_mb: <= 1 MB pulsar_entry_size_le_overflow: > 1 MB
pulsar_delayed_message_index_size_bytes	Gauge	The estimated memory size of DelayedDeliveryTracker for the namespace owned by this broker (in bytes). The in-memory tracker estimates its bitmap index using serialized size; this is not a measurement of total JVM heap usage.
pulsar_delayed_message_index_bucket_total	Gauge	The number of delayed message index buckets (immutable buckets + LastMutableBucket )
pulsar_delayed_message_index_loaded	Gauge	The total number of delayed message indexes for in the memory.
pulsar_delayed_message_index_bucket_snapshot_size_bytes	Gauge	The total size of delayed message index bucket snapshot (in bytes).
pulsar_delayed_message_index_bucket_op_count	Counter	The total number of operation delayed message index bucket snapshots. The state label can be succeed,failed, andall (all means the total number of all states) and the type label can be create,load,delete, and merge.
pulsar_delayed_message_index_bucket_op_latency_ms	Histogram	The latency of delayed message index bucket snapshot operation with a given quantile (threshold). The type label can be create,load,delete, and merge The label quantile can be: quantile="50" is operation latency between (0ms, 50ms] quantile="100" is operation latency between (50ms, 100ms] quantile="500" is operation latency between (100ms, 500ms] quantile="1000" is operation latency between (500ms, 1s] quantile="5000" is operation latency between (1s, 5s] quantile="30000" is operation latency between (5s, 30s] quantile="60000" is operation latency between (30s, 60s] quantile="overflow" is operation latency > 1m

Topic metrics

Topic metrics are only exposed when exposeTopicLevelMetricsInPrometheus is set to true.

All the topic metrics are labeled with the following labels:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you configured in broker.conf.
• namespace: namespace=\$\{pulsar_namespace\}. \$\{pulsar_namespace\} is the namespace name.
• topic: topic=\$\{pulsar_topic\}. \$\{pulsar_topic\} is the topic name.

Name	Type	Description
pulsar_subscriptions_count	Gauge	The number of Pulsar subscriptions of the topic served by this broker.
pulsar_producers_count	Gauge	The number of active producers of the topic connected to this broker.
pulsar_consumers_count	Gauge	The number of active consumers of the topic connected to this broker.
pulsar_rate_in	Gauge	The total message rate of the topic coming into this broker (message per second).
pulsar_rate_out	Gauge	The total message rate of the topic going out from this broker (message per second).
pulsar_publish_rate_limit_times	Gauge	The number of times the publish rate limit is triggered.
pulsar_throughput_in	Gauge	The total throughput of the topic coming into this broker (byte per second).
pulsar_throughput_out	Gauge	The total throughput of the topic going out from this broker (byte per second).
pulsar_consumer_msg_ack_rate	Gauge	The total message acknowledgment rate of the topic connected to this broker (message per second).
pulsar_storage_size	Gauge	The total storage size of the topics in this topic owned by this broker (bytes).
pulsar_storage_logical_size	Gauge	The storage size of topics in the namespace owned by the broker without replicas (in bytes).
pulsar_storage_backlog_size	Gauge	The total backlog size of the topics of this topic owned by this broker (in bytes).
pulsar_storage_offloaded_size	Gauge	The total amount of the data in this topic offloaded to the tiered storage (bytes).
pulsar_storage_backlog_quota_limit	Gauge	The total amount of the data in this topic that limit the backlog quota (bytes).
pulsar_storage_backlog_age_seconds	Gauge	The age of the oldest unacknowledged message (backlog).
pulsar_storage_backlog_quota_exceeded_evictions_total	Counter	The number of times a backlog was evicted since it has exceeded its quota. Includes label quota_type = (time | size)
pulsar_storage_write_rate	Gauge	The total message batches (entries) written to the storage for this topic (message batch per second).
pulsar_storage_read_rate	Gauge	The total message batches (entries) read from the storage for this topic (message batch per second).
pulsar_subscription_delayed	Gauge	The total message batches (entries) are delayed for dispatching.
pulsar_storage_write_latency_le_*	Histogram	The entry rate of a topic that the storage write latency is smaller with a given threshold. Available thresholds: pulsar_storage_write_latency_le_0_5: <= 0.5ms pulsar_storage_write_latency_le_1: <= 1ms pulsar_storage_write_latency_le_5: <= 5ms pulsar_storage_write_latency_le_10: <= 10ms pulsar_storage_write_latency_le_20: <= 20ms pulsar_storage_write_latency_le_50: <= 50ms pulsar_storage_write_latency_le_100: <= 100ms pulsar_storage_write_latency_le_200: <= 200ms pulsar_storage_write_latency_le_1000: <= 1s pulsar_storage_write_latency_le_overflow: > 1s
pulsar_entry_size_le_*	Histogram	The entry rate of a topic that the entry size is smaller with a given threshold. Available thresholds: pulsar_entry_size_le_128: <= 128 bytes pulsar_entry_size_le_512: <= 512 bytes pulsar_entry_size_le_1_kb: <= 1 KB pulsar_entry_size_le_2_kb: <= 2 KB pulsar_entry_size_le_4_kb: <= 4 KB pulsar_entry_size_le_16_kb: <= 16 KB pulsar_entry_size_le_100_kb: <= 100 KB pulsar_entry_size_le_1_mb: <= 1 MB pulsar_entry_size_le_overflow: > 1 MB
pulsar_in_bytes_total	Counter	The total number of messages in bytes received for this topic.
pulsar_in_messages_total	Counter	The total number of messages received for this topic.
pulsar_out_bytes_total	Counter	The total number of messages in bytes read from this topic.
pulsar_out_messages_total	Counter	The total number of messages read from this topic.
pulsar_compaction_removed_event_count	Gauge	The total number of removed events of the compaction.
pulsar_compaction_succeed_count	Gauge	The total number of successes of the compaction.
pulsar_compaction_failed_count	Gauge	The total number of failures of the compaction.
pulsar_compaction_duration_time_in_mills	Gauge	The duration time of the compaction.
pulsar_compaction_read_throughput	Gauge	The read throughput of the compaction.
pulsar_compaction_write_throughput	Gauge	The write throughput of the compaction.
pulsar_compaction_latency_le_*	Histogram	The compaction latency with given quantile. Available thresholds: pulsar_compaction_latency_le_0_5: <= 0.5ms pulsar_compaction_latency_le_1: <= 1ms pulsar_compaction_latency_le_5: <= 5ms pulsar_compaction_latency_le_10: <= 10ms pulsar_compaction_latency_le_20: <= 20ms pulsar_compaction_latency_le_50: <= 50ms pulsar_compaction_latency_le_100: <= 100ms pulsar_compaction_latency_le_200: <= 200ms pulsar_compaction_latency_le_1000: <= 1s pulsar_compaction_latency_le_overflow: > 1s
pulsar_compaction_compacted_entries_count	Gauge	The total number of the compacted entries.
pulsar_compaction_compacted_entries_size	Gauge	The total size of the compacted entries.
pulsar_txn_tb_active_total	Gauge	The number of active transactions on this topic.
pulsar_txn_tb_aborted_total	Counter	The number of aborted transactions on the topic.
pulsar_txn_tb_committed_total	Counter	The number of committed transactions on the topic.
pulsar_delayed_message_index_size_bytes	Gauge	The estimated memory size of DelayedDeliveryTracker for the topic owned by this broker (in bytes). The in-memory tracker estimates its bitmap index using serialized size; this is not a measurement of total JVM heap usage.
pulsar_delayed_message_index_bucket_total	Gauge	The number of delayed message index buckets (immutable buckets + LastMutableBucket )
pulsar_delayed_message_index_loaded	Gauge	The total number of delayed message indexes for in the memory.
pulsar_delayed_message_index_bucket_snapshot_size_bytes	Gauge	The total size of delayed message index bucket snapshot (in bytes).
pulsar_delayed_message_index_bucket_op_count	Counter	The total number of operation delayed message index bucket snapshots. The state label can be succeed,failed, andall (all means the total number of all states) and the type label can be create,load,delete, and merge.
pulsar_delayed_message_index_bucket_op_latency_ms	Histogram	The latency of delayed message index bucket snapshot operation with a given quantile (threshold). The labeltype label can be create,load,delete, and merge The label quantile can be: quantile="50" is operation latency between (0ms, 50ms] quantile="100" is operation latency between (50ms, 100ms] quantile="500" is operation latency between (100ms, 500ms] quantile="1000" is operation latency between (500ms, 1s] quantile="5000" is operation latency between (1s, 5s] quantile="30000" is operation latency between (5s, 30s] quantile="60000" is operation latency between (30s, 60s] quantile="overflow" is operation latency > 1m

Producer metrics

Enable exposeProducerLevelMetricsInPrometheus and topic-level metrics to expose producer metrics. Labels include cluster, namespace, topic, producer_name, and producer_id.

Name	Type	Description
pulsar_producer_msg_rate_in	Gauge	Messages published per second by this producer.
pulsar_producer_msg_throughput_in	Gauge	Bytes published per second by this producer.
pulsar_producer_msg_average_Size	Gauge	Average published message size in bytes. The capital S is part of the exported name.

Replication metrics

If a namespace is configured to be replicated among multiple Pulsar clusters, the corresponding replication metrics is also exposed when replicationMetricsEnabled is enabled.

All the replication metrics are also labelled with remoteCluster=\$\{pulsar_remote_cluster\}.

Name	Type	Description
pulsar_replication_rate_in	Gauge	The total message rate of the namespace replicating from remote cluster (messages/second).
pulsar_replication_rate_out	Gauge	The total message rate of the namespace replicating to remote cluster (messages/second).
pulsar_replication_throughput_in	Gauge	The total throughput of the namespace replicating from remote cluster (bytes/second).
pulsar_replication_throughput_out	Gauge	The total throughput of the namespace replicating to remote cluster (bytes/second).
pulsar_replication_backlog	Gauge	The total backlog of the namespace replicating to remote cluster (messages).
pulsar_replication_rate_expired	Gauge	Total rate of messages expired (messages/second).
pulsar_replication_connected_count	Gauge	The count of replication-subscriber up and running to replicate to remote cluster.
pulsar_replication_delay_in_seconds	Gauge	Time in seconds from the time a message was produced to the time when it is about to be replicated.

Replicated-subscription snapshots

Snapshot metrics describe cross-cluster replicated-subscription position tracking.

Name	Type	Description
pulsar_replicated_subscriptions_pending_snapshots	Gauge	Counter of currently pending snapshots
pulsar_replicated_subscriptions_snapshot_ms	Summary	Time taken to create a consistent snapshot across clusters
pulsar_replicated_subscriptions_timedout_snapshots_total	Counter	Counter of timed out snapshots
pulsar_replication_disconnected_count	Gauge	Number of disconnected replicators, labeled by namespace and remote cluster.

Topic lookup metrics

Name	Type	Description
pulsar_broker_load_manager_bundle_assignment	Gauge	The summary of latency of bundles ownership operations.
pulsar_broker_lookup	Gauge	The latency of all lookup operations.
pulsar_broker_lookup_redirects	Gauge	The number of lookup redirected requests.
pulsar_broker_lookup_answers	Gauge	The number of lookup responses (i.e. not redirected requests).
pulsar_broker_lookup_failures	Gauge	The number of lookup failures.
pulsar_broker_lookup_pending_requests	Gauge	The number of pending lookups in broker. When it is up to the threshold, new requests are rejected.
pulsar_broker_topic_load_pending_requests	Gauge	The load of pending topic operations.

Topic loading failures

Name	Type	Description
topic_load_failed_total	Counter	Persistent topic load failures, labelled by reason.
pulsar_topic_load_failed_count	Gauge	Compatibility total across all failure reasons. This is brk_topic_load_failed_count in the broker metrics API.

Each persistent topic-loading operation records at most one failure. The reason values are bundle_unloading, failed_load_namespace_policies, failed_load_topic_policies, failed_load_ml, failed_check_ownership, failed_access_metadata_store, failed_init, timeout, timeout_load_namespace_policies, timeout_load_topic_policies, timeout_load_ml, timeout_init, timeout_dedup, and others. A timeout identifies the pending loading stage when known; timeout is the fallback. All reason series are initialized to zero.

Aggregate over reason for an overall failure rate, or select a reason to diagnose a particular loading stage. For example:

sum without (reason) (rate(topic_load_failed_total[5m]))

Storage and compaction bucket companions

The pulsar_storage_write_latency_le_* and pulsar_storage_ledger_write_latency_le_* series contain per-period latency bucket counts. The ledger-write family isolates BookKeeper add-entry latency. Bucket names express upper bounds in milliseconds, while the underlying measurements and _sum values use microseconds. Each family also exports _count, _sum, and _overflow; for example, pulsar_storage_ledger_write_latency_count, pulsar_storage_ledger_write_latency_sum, and pulsar_storage_ledger_write_latency_overflow.

The pulsar_compaction_latency_le_* family follows the same bucket naming and microsecond sum units, with pulsar_compaction_latency_count, pulsar_compaction_latency_sum, and pulsar_compaction_latency_overflow companions. Its stats refresh when collected. The pulsar_entry_size_le_* family measures bytes, with pulsar_entry_size_count and pulsar_entry_size_sum companions. These bucket families are exported as gauges; do not treat them as cumulative Prometheus histograms.

ManagedLedger metrics

All the managedLedger metrics are labeled with the following labels:

• cluster: cluster=${pulsar_cluster}. ${pulsar_cluster} is the cluster name that you have configured in the broker.conf file.
• namespace: namespace=${pulsar_namespace}. ${pulsar_namespace} is the namespace name.
• quantile: quantile=${quantile}. Quantile is only for Histogram type metric, and represents the threshold for given Buckets.

Name	Type	Description
pulsar_ml_AddEntryBytesRate	Gauge	The B/s rate of messages added
pulsar_ml_AddEntryWithReplicasBytesRate	Gauge	The B/s rate of messages added with replicas
pulsar_ml_AddEntryErrors	Gauge	The number of addEntry requests that failed
pulsar_ml_AddEntryLatencyBuckets	Histogram	The latency of adding a ledger entry with a given quantile (threshold), including time spent waiting in queue on the broker side. Available quantile: quantile="0.0_0.5" is AddEntryLatency between (0.0ms, 0.5ms] quantile="0.5_1.0" is AddEntryLatency between (0.5ms, 1.0ms] quantile="1.0_5.0" is AddEntryLatency between (1ms, 5ms] quantile="5.0_10.0" is AddEntryLatency between (5ms, 10ms] quantile="10.0_20.0" is AddEntryLatency between (10ms, 20ms] quantile="20.0_50.0" is AddEntryLatency between (20ms, 50ms] quantile="50.0_100.0" is AddEntryLatency between (50ms, 100ms] quantile="100.0_200.0" is AddEntryLatency between (100ms, 200ms] quantile="200.0_1000.0" is AddEntryLatency between (200ms, 1s]
pulsar_ml_AddEntryLatencyBuckets_OVERFLOW	Gauge	The number of times the AddEntryLatency is longer than 1 second
pulsar_ml_AddEntryMessagesRate	Gauge	The msg/s rate of messages added
pulsar_ml_AddEntrySucceed	Gauge	The number of addEntry requests that succeeded
pulsar_ml_EntrySizeBuckets	Histogram	The added entry size of a ledger with a given quantile. Available quantile: quantile="0.0_128.0" is EntrySize between (0byte, 128byte] quantile="128.0_512.0" is EntrySize between (128byte, 512byte] quantile="512.0_1024.0" is EntrySize between (512byte, 1KB] quantile="1024.0_2048.0" is EntrySize between (1KB, 2KB] quantile="2048.0_4096.0" is EntrySize between (2KB, 4KB] quantile="4096.0_16384.0" is EntrySize between (4KB, 16KB] quantile="16384.0_102400.0" is EntrySize between (16KB, 100KB] quantile="102400.0_1232896.0" is EntrySize between (100KB, 1MB]
pulsar_ml_EntrySizeBuckets_OVERFLOW	Gauge	The number of times the EntrySize is larger than 1MB
pulsar_ml_LedgerSwitchLatencyBuckets	Histogram	The ledger switch latency with a given quantile. Available quantile: quantile="0.0_0.5" is EntrySize between (0ms, 0.5ms] quantile="0.5_1.0" is EntrySize between (0.5ms, 1ms] quantile="1.0_5.0" is EntrySize between (1ms, 5ms] quantile="5.0_10.0" is EntrySize between (5ms, 10ms] quantile="10.0_20.0" is EntrySize between (10ms, 20ms] quantile="20.0_50.0" is EntrySize between (20ms, 50ms] quantile="50.0_100.0" is EntrySize between (50ms, 100ms] quantile="100.0_200.0" is EntrySize between (100ms, 200ms] quantile="200.0_1000.0" is EntrySize between (200ms, 1000ms]
pulsar_ml_LedgerSwitchLatencyBuckets_OVERFLOW	Gauge	The number of times the ledger switch latency is longer than 1 second
pulsar_ml_LedgerAddEntryLatencyBuckets	Histogram	The latency for bookie client to persist a ledger entry from broker to BookKeeper service with a given quantile (threshold). Available quantile: quantile="0.0_0.5" is LedgerAddEntryLatency between (0.0ms, 0.5ms] quantile="0.5_1.0" is LedgerAddEntryLatency between (0.5ms, 1.0ms] quantile="1.0_5.0" is LedgerAddEntryLatency between (1ms, 5ms] quantile="5.0_10.0" is LedgerAddEntryLatency between (5ms, 10ms] quantile="10.0_20.0" is LedgerAddEntryLatency between (10ms, 20ms] quantile="20.0_50.0" is LedgerAddEntryLatency between (20ms, 50ms] quantile="50.0_100.0" is LedgerAddEntryLatency between (50ms, 100ms] quantile="100.0_200.0" is LedgerAddEntryLatency between (100ms, 200ms] quantile="200.0_1000.0" is LedgerAddEntryLatency between (200ms, 1s]
pulsar_ml_LedgerAddEntryLatencyBuckets_OVERFLOW	Gauge	The number of times the LedgerAddEntryLatency is longer than 1 second
pulsar_ml_MarkDeleteRate	Gauge	The rate of mark-delete ops/s
pulsar_ml_NumberOfMessagesInBacklog	Gauge	The number of backlog messages for all the consumers
pulsar_ml_ReadEntriesBytesRate	Gauge	The B/s rate of messages read
pulsar_ml_ReadEntriesErrors	Gauge	The number of readEntries requests that failed
pulsar_ml_ReadEntriesRate	Gauge	The msg/s rate of messages read
pulsar_ml_ReadEntriesSucceeded	Gauge	The number of readEntries requests that succeeded
pulsar_ml_StoredMessagesSize	Gauge	The total size of the messages in active ledgers (accounting for the multiple copies stored)

Managed-ledger read cache and permits

Pending-read counters describe reuse of concurrent BookKeeper reads. In-flight byte gauges describe the read limiter and report -1 when limiting is disabled. These are broker-wide metrics.

Name	Type	Description
pulsar_ml_cache_pendingreads_entries_notread_total	Counter	Total number of entries not read from BK
pulsar_ml_cache_pendingreads_entries_read_total	Counter	Total number of entries read from BK
pulsar_ml_cache_pendingreads_matched_total	Counter	Pending reads reused with perfect range match
pulsar_ml_cache_pendingreads_matched_included_total	Counter	Pending reads reused by attaching to a read with a larger range
pulsar_ml_cache_pendingreads_matched_overlapping_miss_both_total	Counter	Pending reads that didn't find a match but they partially overlap with another read
pulsar_ml_cache_pendingreads_matched_overlapping_miss_left_total	Counter	Pending reads that didn't find a match but they partially overlap with another read
pulsar_ml_cache_pendingreads_matched_overlapping_miss_right_total	Counter	Pending reads that didn't find a match but they partially overlap with another read
pulsar_ml_cache_pendingreads_missed_total	Counter	Pending reads that didn't find a match
pulsar_ml_reads_available_inflight_bytes	Gauge	Available space for inflight data read from storage or cache
pulsar_ml_reads_inflight_bytes	Gauge	Estimated number of bytes retained by data read from storage or cache
pulsar_ml_ReadEntriesOpsCacheMissesRate	Gauge	Managed-ledger read operations per second that miss the cache, aggregated by namespace when exposeManagedLedgerMetricsInPrometheus=true.

Managed cursor acknowledgment state

The acknowledgment state is persistent to the ledger first. When the acknowledgment state fails to be persistent to the ledger, they are persistent to ZooKeeper. To track the stats of acknowledgment, you can configure the metrics for the managed cursor.

All the cursor acknowledgment state metrics are labeled with the following labels:

• namespace: namespace=\$\{pulsar_namespace\}. \$\{pulsar_namespace\} is the namespace name.

• ledger_name: ledger_name=\$\{pulsar_ledger_name\}. \$\{pulsar_ledger_name\} is the ledger name.

• cursor_name: ledger_name=\$\{pulsar_cursor_name\}. \$\{pulsar_cursor_name\} is the cursor name.

Name	Type	Description
pulsar_ml_cursor_persistLedgerSucceed	Gauge	The number of acknowledgment states that is persistent to a ledger.
pulsar_ml_cursor_persistLedgerErrors	Gauge	The number of ledger errors occurred when acknowledgment states fail to be persistent to the ledger.
pulsar_ml_cursor_persistZookeeperSucceed	Gauge	The number of acknowledgment states that is persistent to ZooKeeper.
pulsar_ml_cursor_persistZookeeperErrors	Gauge	The number of ledger errors occurred when acknowledgment states fail to be persistent to ZooKeeper.
pulsar_ml_cursor_nonContiguousDeletedMessagesRange	Gauge	The number of non-contiguous deleted messages ranges.
pulsar_ml_cursor_writeLedgerSize	Gauge	The size of write to ledger.
pulsar_ml_cursor_writeLedgerLogicalSize	Gauge	The size of write to ledger (accounting for without replicas).
pulsar_ml_cursor_readLedgerSize	Gauge	The size of read from ledger.

Subscription metrics

Subscription metrics are only exposed when exposeTopicLevelMetricsInPrometheus is set to true.

To expose subscription backlog age, also set exposeSubscriptionBacklogAgeInPrometheus=true in broker.conf and restart the broker. It defaults to false. When disabled, the broker skips computing subscription backlog age and the topic-stat field oldestBacklogMessageAgeSeconds is -1; the Prometheus age series is absent. See Monitor subscription backlog age.

All the subscription metrics are labeled with the following labels:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you have configured in the broker.conf file.
• namespace: namespace=\$\{pulsar_namespace\}. \$\{pulsar_namespace\} is the namespace name.
• topic: topic=\$\{pulsar_topic\}. \$\{pulsar_topic\} is the topic name.
• subscription: subscription=\$\{subscription\}. \$\{subscription\} is the topic subscription name.

Name	Type	Description
pulsar_subscription_back_log	Gauge	The number of entries (messages/batched-messages) in unacknowledged state for a subscription
pulsar_subscription_back_log_no_delayed	Gauge	The backlog of a subscription that does not contain the delay messages (entries).
pulsar_subscription_storage_backlog_age_seconds	Gauge	Best-effort age, in seconds, of the oldest unacknowledged message in the subscription storage backlog. Requires exposeSubscriptionBacklogAgeInPrometheus=true. Omitted while the age is unavailable.
pulsar_subscription_delayed	Gauge	The total number of messages are delayed to be dispatched for a subscription (messages).
pulsar_subscription_msg_rate_redeliver	Gauge	The total message rate for message being redelivered (message per second).
pulsar_subscription_unacked_messages	Gauge	The number of entries (messages/batched-messages) dispatched to consumers and are still unacknowledged
pulsar_subscription_blocked_on_unacked_messages	Gauge	Indicate whether a subscription is blocked on unacknowledged messages or not. 1 means the subscription is blocked on waiting for unacknowledged messages to be acked. 0 means the subscription is not blocked on waiting for unacknowledged messages to be acked.
pulsar_subscription_msg_rate_out	Gauge	The total message dispatch rate for a subscription (message per second).
pulsar_subscription_msg_throughput_out	Gauge	The total message dispatch throughput for a subscription (byte per second).
pulsar_subscription_msg_ack_rate	Gauge	The total message acknowledgment rate for a subscription (message per second).
pulsar_subscription_last_expire_timestamp	Gauge	The expiration timestamp of the last message on the subscription.
pulsar_subscription_last_acked_timestamp	Gauge	The timestamp of the last acknowledged message on the subscription.
pulsar_subscription_last_consumed_flow_timestamp	Gauge	The timestamp of the last received consume flow command on the subscription.
pulsar_subscription_last_consumed_timestamp	Gauge	The timestamp of the last consumed message on the subscription.
pulsar_subscription_last_mark_delete_advanced_timestamp	Gauge	The advanced timestamp of the last MarkDelete position on the subscription.
pulsar_subscription_msg_rate_expired	Gauge	The total rate of messages expired on this subscription (message per second).
pulsar_subscription_total_msg_expired	Gauge	The total number of messages expired on this subscription.
pulsar_subscription_msg_drop_rate	Gauge	The rate of messages dropped on this subscription (message per second).
pulsar_subscription_consumers_count	Gauge	The number of connected consumers on this subscription.
pulsar_subscription_filter_processed_msg_count	Counter	The number of messages processed by EntryFilter.
pulsar_subscription_filter_accepted_msg_count	Counter	The number of messages accepted by EntryFilter.
pulsar_subscription_filter_rejected_msg_count	Counter	The number of messages rejected by EntryFilter.
pulsar_subscription_filter_rescheduled_msg_count	Counter	The number of messages rescheduled by EntryFilter.
pulsar_delayed_message_index_size_bytes	Gauge	The estimated memory size of DelayedDeliveryTracker for the subscription owned by this broker (in bytes). The in-memory tracker estimates its bitmap index using serialized size; this is not a measurement of total JVM heap usage.
pulsar_delayed_message_index_bucket_total	Gauge	The number of delayed message index buckets (immutable buckets + LastMutableBucket )
pulsar_delayed_message_index_loaded	Gauge	The total number of delayed message indexes for in the memory.
pulsar_delayed_message_index_bucket_snapshot_size_bytes	Gauge	The total size of delayed message index bucket snapshot (in bytes).
pulsar_delayed_message_index_bucket_op_count	Counter	The total number of operation delayed message index bucket snapshots. The state label can be succeed,failed, andall (all means the total number of all states) and the type label can be create,load,delete, and merge.
pulsar_delayed_message_index_bucket_op_latency_ms	Histogram	The latency of delayed message index bucket snapshot operation with a given quantile (threshold). The labeltype label can be create,load,delete, and merge The label quantile can be: quantile="50" is operation latency between (0ms, 50ms] quantile="100" is operation latency between (50ms, 100ms] quantile="500" is operation latency between (100ms, 500ms] quantile="1000" is operation latency between (500ms, 1s] quantile="5000" is operation latency between (1s, 5s] quantile="30000" is operation latency between (5s, 30s] quantile="60000" is operation latency between (30s, 60s] quantile="overflow" is operation latency > 1m

Additional topic and subscription metrics

These series use the topic/subscription labels of their corresponding sections. Topic-level series require exposeTopicLevelMetricsInPrometheus=true.

Name	Type	Description
pulsar_average_msg_size	Gauge	Average incoming message size in bytes for a topic.
pulsar_msg_backlog	Gauge	Backlog summed across subscriptions; emitted at the configured namespace or topic aggregation level.
pulsar_storage_backlog_quota_limit_time	Gauge	Topic backlog age quota in seconds.
pulsar_storage_read_cache_misses_rate	Gauge	Read operations per second missing the storage cache, at namespace or topic level.
pulsar_subscription_delayed_message_index_size_bytes	Gauge	Memory used by the subscription delayed-message index, in bytes.
pulsar_subscription_dispatch_throttled_msg_events	Gauge	Cumulative message-rate throttling events. The reason label is subscription, topic, or broker.
pulsar_subscription_dispatch_throttled_bytes_events	Gauge	Cumulative byte-rate throttling events. The reason label is subscription, topic, or broker.
pulsar_subscription_in_replay	Gauge	Entries in subscription replay queues, exposed per subscription with topic-level metrics or aggregated by namespace.

Consumer metrics

Consumer metrics are only exposed when both exposeTopicLevelMetricsInPrometheus and exposeConsumerLevelMetricsInPrometheus are set to true.

All the consumer metrics are labeled with the following labels:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you have configured in the broker.conf file.
• namespace: namespace=\$\{pulsar_namespace\}. \$\{pulsar_namespace\} is the namespace name.
• topic: topic=\$\{pulsar_topic\}. \$\{pulsar_topic\} is the topic name.
• subscription: subscription=\$\{subscription\}. \$\{subscription\} is the topic subscription name.
• consumer_name: consumer_name=\$\{consumer_name\}. \$\{consumer_name\} is the topic consumer name.
• consumer_id: consumer_id=\$\{consumer_id\}. \$\{consumer_id\} is the topic consumer id.

Name	Type	Description
pulsar_consumer_msg_rate_redeliver	Gauge	The total message rate for message being redelivered (message per second).
pulsar_consumer_unacked_messages	Gauge	The total number of unacknowledged messages of a consumer (messages).
pulsar_consumer_blocked_on_unacked_messages	Gauge	Indicate whether a consumer is blocked on unacknowledged messages or not. 1 means the consumer is blocked on waiting for unacknowledged messages to be acked. 0 means the consumer is not blocked on waiting for unacknowledged messages to be acked.
pulsar_consumer_msg_rate_out	Gauge	The total message dispatch rate for a consumer (message per second).
pulsar_consumer_msg_ack_rate	Gauge	The total rate of message ack (message per second).
pulsar_consumer_msg_throughput_out	Gauge	The total message dispatch throughput for a consumer (byte per second).
pulsar_consumer_available_permits	Gauge	The available permits for for a consumer.

Managed ledger bookie client metrics

All the managed ledger bookie client metrics are labeled with the following labels:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you have configured in the broker.conf file.

Name	Type	Description
pulsar_managedLedger_client_bookkeeper_ml_scheduler_completed_tasks_*	Gauge	The number of tasks the scheduler executor execute completed. The number of metrics determined by the scheduler executor thread number configured by managedLedgerNumSchedulerThreads in broker.conf.
pulsar_managedLedger_client_bookkeeper_ml_scheduler_queue_*	Gauge	The number of tasks queued in the scheduler executor's queue. The number of metrics determined by the scheduler executor's thread number configured by managedLedgerNumSchedulerThreads in broker.conf.
pulsar_managedLedger_client_bookkeeper_ml_scheduler_total_tasks_*	Gauge	The total number of tasks the scheduler executor received. The number of metrics determined by the scheduler executor's thread number configured by managedLedgerNumSchedulerThreads in broker.conf.
pulsar_managedLedger_client_bookkeeper_ml_scheduler_task_execution	Summary	The scheduler task execution latency calculated in milliseconds.
pulsar_managedLedger_client_bookkeeper_ml_scheduler_task_queued	Summary	The scheduler task queued latency calculated in milliseconds.

Token metrics

All the token metrics are labeled with the following labels:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you have configured in the broker.conf file.

Name	Type	Description
pulsar_expired_token_total	Counter	The number of expired tokens in Pulsar.
pulsar_expiring_token_minutes	Histogram	The remaining time of expiring tokens in minutes.

Authentication metrics

All the authentication metrics are labeled with the following labels:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you have configured in the broker.conf file.
• provider_name: provider_name=\$\{provider_name\}. \$\{provider_name\} is the class name of the authentication provider.
• auth_method: auth_method=\$\{auth_method\}. \$\{auth_method\} is the authentication method of the authentication provider.
• reason: reason=\$\{reason\}. \$\{reason\} is the reason for failing authentication operation. (This label is only for pulsar_authentication_failures_total.)

Name	Type	Description
pulsar_authentication_success_total	Counter	The number of successful authentication operations.
pulsar_authentication_failures_total	Counter	The number of failing authentication operations.

Connection metrics

All the connection metrics are labelled with the following labels:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you have configured in the broker.conf file.
• broker: broker=\$\{advertised_address\}. \$\{advertised_address\} is the advertised address of the broker.
• metric: metric=\$\{metric\}. \$\{metric\} is the connection metric collective name.

Name	Type	Description
pulsar_active_connections	Gauge	The number of active connections.
pulsar_connection_created_total_count	Gauge	The total number of connections.
pulsar_connection_create_success_count	Gauge	The number of successfully created connections.
pulsar_connection_create_fail_count	Gauge	The number of failed connections.
pulsar_connection_closed_total_count	Gauge	The total number of closed connections.
pulsar_broker_throttled_connections	Gauge	The number of throttled connections.

Jetty metrics

For a functions-worker running separately from brokers, its Jetty metrics are only exposed when includeStandardPrometheusMetrics is set to true.

All the jetty metrics are labeled with the following labels:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you have configured in the broker.conf file.

Name	Type	Description
jetty_requests_total	Counter	Number of requests.
jetty_requests_active	Gauge	Number of requests currently active.
jetty_requests_active_max	Gauge	Maximum number of requests that have been active at once.
jetty_request_time_max_seconds	Gauge	Maximum time spent handling requests.
jetty_request_time_seconds_total	Counter	Total time spent in all request handling.
jetty_dispatched_total	Counter	Number of dispatches.
jetty_dispatched_active	Gauge	Number of dispatches currently active.
jetty_dispatched_active_max	Gauge	Maximum number of active dispatches being handled.
jetty_dispatched_time_max	Gauge	Maximum time spent in dispatch handling.
jetty_dispatched_time_seconds_total	Counter	Total time spent in dispatch handling.
jetty_async_requests_total	Counter	Total number of async requests.
jetty_async_requests_waiting	Gauge	Currently waiting async requests.
jetty_async_requests_waiting_max	Gauge	Maximum number of waiting async requests.
jetty_async_dispatches_total	Counter	Number of requests that have been asynchronously dispatched.
jetty_expires_total	Counter	Number of async requests that have expired.
jetty_responses_total	Counter	Number of responses, labeled by status code. The code label can be "1xx", "2xx", "3xx", "4xx", or "5xx".
jetty_stats_seconds	Gauge	Time in seconds stats have been collected for.
jetty_responses_bytes_total	Counter	Total number of bytes across all responses.

Schema metrics

For a functions-worker running separately from brokers, its schema metrics are only exposed when includeStandardPrometheusMetrics is set to true.

All the schema metrics are labeled with the following labels:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you have configured in the broker.conf file.

Name	Type	Description
pulsar_schema_del_ops_failed_total	Counter	Number of failed operations to delete schemas.
pulsar_schema_get_ops_failed_total	Counter	Number of failed operations to get schemas.
pulsar_schema_put_ops_failed_total	Counter	Number of failed operations to send schemas.
pulsar_schema_compatible_total	Counter	Number of compatible schemas.
pulsar_schema_incompatible_total	Counter	Number of incompatible schemas.
pulsar_schema_del_ops_latency	Summary	Latency of successful operations to delete schemas.
pulsar_schema_get_ops_latency	Summary	Latency of successful operations to get schemas.
pulsar_schema_put_ops_latency	Summary	Latency of successful operations to send schemas.

Offload metrics

Offloader collectors retain their brk_ prefix: they are registered directly in the Prometheus registry. The brk_ to pulsar_ conversion applies to generated broker metrics, not these collectors. Topic labels are present only when topic-level offloader metrics are enabled.

All the offload metrics are labeled with the following labels:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you configured in broker.conf.
• namespace: namespace=\$\{pulsar_namespace\}. \$\{pulsar_namespace\} is the namespace name.
• topic: topic=\$\{pulsar_topic\}. \$\{pulsar_topic\} is the topic name.

Name	Type	Description
brk_ledgeroffloader_offload_error_total	Counter	The number of failed operations to offload.
brk_ledgeroffloader_offload_rate	Gauge	The rate of offloading(byte per second).
brk_ledgeroffloader_read_offload_error_total	Counter	The number of failed operations to read offload ledgers.
brk_ledgeroffloader_read_offload_rate	Gauge	The rate of reading entries from offload ledgers(byte per second).
brk_ledgeroffloader_write_storage_error_total	Counter	The number of failed operations to write to storage.
brk_ledgeroffloader_read_offload_index_latency	Summary	The latency of reading index from offload ledgers, in microseconds.
brk_ledgeroffloader_read_offload_data_latency	Summary	The latency of reading data from offload ledgers, in microseconds.
brk_ledgeroffloader_read_ledger_latency	Summary	The latency of reading entries from BookKeeper, in microseconds.
brk_ledgeroffloader_delete_offload_ops_total	Counter	The total number of successful and failed operations to delete offload ledgers, distinguished by the status label.

Offload read bytes

Offloader collectors retain the brk_ prefix on the Prometheus endpoint.

Name	Type	Description
brk_ledgeroffloader_read_bytes_total	Counter	Total bytes read from offloaded storage. Uses the same namespace/topic labels as the other offloader collectors.

Web service executor metrics

For functions workers running separately from brokers, their Jetty metrics are only exposed when includeStandardPrometheusMetrics is set to true.

All the web service executor metrics are labeled with the following labels:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you have configured in the broker.conf file.

Name	Type	Description
pulsar_web_executor_max_threads	GAUGE	The max threads of pulsar-web thread pool
pulsar_web_executor_min_threads	GAUGE	The min threads of pulsar-web thread pool
pulsar_web_executor_idle_threads	GAUGE	The idle threads of pulsar-web thread pool
pulsar_web_executor_active_threads	GAUGE	The number of threads performing tasks of pulsar-web thread pool
pulsar_web_executor_current_threads	GAUGE	The number of threads in the pulsar-web thread pool

Metadata store metrics

All the metadata store metrics are labeled with the following labels:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you configured in broker.conf.
• name: name=\$\{metadata-store\|configuration-metadata-store\|state-metadata-store\}. \$\{name\} is the metadata store name.

Name	Type	Description
pulsar_metadata_store_ops_latency_ms	Histogram	The latency of getting/deleting/putting data from/to metadata store, in milliseconds.
pulsar_metadata_store_put_bytes_total	Counter	The number of data put to metadata store.
pulsar_batch_metadata_store_executor_queue_size	Gauge	The number of blocking operations in metadata store executor.
pulsar_batch_metadata_store_queue_wait_time_ms	Histogram	The waiting time of batch operations.
pulsar_batch_metadata_store_batch_execute_time_ms	Histogram	The duration of the batch execution in milliseconds.
pulsar_batch_metadata_store_batch_size	Histogram	The number of read/write operations in the batch.

JVM Metrics

Process Metrics

Name	Type	Description
process_cpu_seconds_total	Counter	Total user and system CPU time spent in seconds.
process_start_time_seconds	Gauge	Start time of the process since unix epoch in seconds.
process_open_fds	Gauge	Number of open file descriptors.
process_max_fds	Gauge	Maximum number of open file descriptors.
process_virtual_memory_bytes	Gauge	Virtual memory size in bytes.
process_resident_memory_bytes	Gauge	Resident memory size in bytes.

Memory Metrics

Name	Type	Description
jvm_memory_objects_pending_finalization	Gauge	The number of objects waiting in the finalizer queue.
jvm_memory_bytes_used	Gauge	Used bytes of a given JVM memory area.
jvm_memory_bytes_committed	Gauge	Committed (bytes) of a given JVM memory area.
jvm_memory_bytes_max	Gauge	Max (bytes) of a given JVM memory area.
jvm_memory_bytes_init	Gauge	Initial bytes of a given JVM memory area.
jvm_memory_pool_bytes_used	Used bytes of a given JVM memory pool.
jvm_memory_pool_bytes_committed	Gauge	Committed bytes of a given JVM memory pool.
jvm_memory_pool_bytes_max	Gauge	Max bytes of a given JVM memory pool.
jvm_memory_pool_bytes_init	Gauge	Initial bytes of a given JVM memory pool.
jvm_memory_pool_collection_used_bytes	Gauge	Used bytes after the last collection of a given JVM memory pool.
jvm_memory_pool_collection_committed_bytes	Gauge	Committed after last collection bytes of a given JVM memory pool.
jvm_memory_pool_collection_max_bytes	Gauge	Max bytes after the last collection of a given JVM memory pool.
jvm_memory_pool_collection_init_bytes	Gauge	Initial after last collection bytes of a given JVM memory pool.
jvm_memory_pool_allocated_bytes_total	Counter	Total bytes allocated in a given JVM memory pool. Only updated after GC, not continuously.

Direct memory metrics

These gauges supplement the JVM memory pool metrics.

Name	Type	Description
jvm_memory_direct_bytes_used	Gauge	Direct memory currently used, in bytes.
jvm_memory_direct_bytes_max	Gauge	Maximum direct memory available to the JVM, in bytes.

Buffer Pools Metrics

Name	Type	Description
jvm_buffer_pool_used_bytes	Gauge	Used bytes of a given JVM buffer pool.
jvm_buffer_pool_capacity_bytes	Gauge	Bytes capacity of a given JVM buffer pool.
jvm_buffer_pool_used_buffers	Gauge	Used buffers of a given JVM buffer pool.

Garbage Collectors Metrics

Name	Type	Description
jvm_gc_collection_seconds	Summary	Time spent in a given JVM garbage collector in seconds.

Threads Metrics

Name	Type	Description
jvm_threads_current	Gauge	Current thread count of a JVM
jvm_threads_daemon	Gauge	Daemon thread count of a JVM
jvm_threads_peak	Gauge	Peak thread count of a JVM
jvm_threads_started_total	Counter	Started thread count of a JVM
jvm_threads_deadlocked	Gauge	Cycles of JVM-threads that are in deadlock waiting to acquire object monitors or ownable synchronizers
jvm_threads_deadlocked_monitor	Gauge	Cycles of JVM-threads that are in deadlock waiting to acquire object monitors
jvm_threads_state	Gauge	Current count of threads by state

Classloaders Metrics

Name	Type	Description
jvm_classes_currently_loaded	Gauge	The number of classes that are currently loaded in the JVM
jvm_classes_loaded_total	Counter	The total number of classes that have been loaded since the JVM has started execution
jvm_classes_unloaded_total	Counter	The total number of classes that have been unloaded since the JVM has started execution

Load balancing

This section shows all metrics related to broker load balancing.

note

• Load balancing metrics are not exposed by default. If you want to access load balancing metrics, you need to expose them by setting the following configurations in the broker.conf or standalone.conf file and ensure that your cluster has an active producer or consumer.

  loadManagerClassName=org.apache.pulsar.broker.loadbalance.impl.ModularLoadManagerImpl
  loadBalancerEnabled=true
  exposeBundlesMetricsInPrometheus=true // Add this configuration to standalone.conf

• Metrics with an asterisk (*) are only available in the extensible load balancer.

LoadBalancing metrics

All the loadbalancing metrics are labeled with the following labels:

• cluster: cluster=${pulsar_cluster}. ${pulsar_cluster} is the cluster name that you have configured in the broker.conf file.
• broker: broker=${broker}. ${broker} is the IP address of the broker
• metric: metric="loadBalancing".

Name	Type	Description
pulsar_lb_bandwidth_in_usage	Gauge	The broker inbound bandwidth usage (in percent).
pulsar_lb_bandwidth_out_usage	Gauge	The broker outbound bandwidth usage (in percent).
pulsar_lb_cpu_usage	Gauge	The broker cpu usage (in percent).
pulsar_lb_directMemory_usage	Gauge	The broker process direct memory usage (in percent).
pulsar_lb_memory_usage	Gauge	The broker process memory usage (in percent).
pulsar_lb_resource_usage {feature=max}*	Gauge	The max resource usage of the bandwidth, CPU, memory, and direct_memory.
pulsar_lb_resource_usage {feature=max_ema}*	Gauge	The broker load score (WeightedMaxEMA).

BundleUnloading metrics

All the bundleUnloading metrics are labeled with the following labels:

• cluster: cluster=${pulsar_cluster}. ${pulsar_cluster} is the cluster name that you have configured in the broker.conf file.
• bundle: bundle=${bundle}. ${bundle} is the bundle range on this broker.
• metric: metric="bundleUnloading".

Name	Type	Description
pulsar_lb_unload_broker_total	Counter	Unload broker count in this bundle unloading
pulsar_lb_unload_bundle_total	Counter	Bundle unload count in this bundle unloading. If the value of pulsar_lb_unload_bundle_total is greater than zero, it means that the bundle has been unloaded.
pulsar_lb_unload_broker_breakdown_total{result, reason}*	Counter	Unload broker breakdown count grouped by result and reason labels.
pulsar_lb_resource_usage_stats{feature=max_ema, stat=avg}*	Gauge	The average of brokers' load scores.
pulsar_lb_resource_usage_stats{feature=max_ema, stat=std}*	Gauge	The standard deviation of brokers' load scores.

BundleSplit metrics

All the bundleUnloading metrics are labeled with the following labels:

• cluster: cluster=${pulsar_cluster}. ${pulsar_cluster} is the cluster name that you have configured in the broker.conf file.
• bundle: bundle=${bundle}. ${bundle} is the bundle range on this broker.
• metric: metric="bundlesSplit".

Name	Type	Description
pulsar_lb_bundles_split_total	Counter	The total count of bundle split in this leader broker
pulsar_lb_bundles_split_breakdown_total{result, reason}*	Counter	Bundle split breakdown count grouped by the result and reason labels.

Bundle metrics

Set exposeBundlesMetricsInPrometheus=true in the broker configuration to expose these metrics; its default is false. The extensible load manager also emits all seven bundle metrics below.

All the bundle metrics are labeled with the following labels:

• cluster: cluster=${pulsar_cluster}. ${pulsar_cluster} is the cluster name that you have configured in the broker.conf file.
• broker: broker=${broker}. ${broker} is the IP address of the broker
• bundle: bundle=${bundle}. ${bundle} is the bundle range on this broker
• metric: metric="bundle".

Name	Type	Description
pulsar_bundle_msg_rate_in	Gauge	The total message rate coming into the topics in this bundle (message per second).
pulsar_bundle_msg_rate_out	Gauge	The total message rate going out from the topics in this bundle (message per second).
pulsar_bundle_topics_count	Gauge	The topic count in this bundle.
pulsar_bundle_consumer_count	Gauge	The consumer count of the topics in this bundle.
pulsar_bundle_producer_count	Gauge	The producer count of the topics in this bundle.
pulsar_bundle_msg_throughput_in	Gauge	The total throughput coming into the topics in this bundle (byte per second).
pulsar_bundle_msg_throughput_out	Gauge	The total throughput going out from the topics in this bundle (byte per second).

Bundle assign metrics

All the bundle assign metrics are labeled with the following labels:

• cluster: cluster=${pulsar_cluster}. ${pulsar_cluster} is the cluster name you have configured in the broker.conf file.
• broker: broker=${broker}. ${broker} is the IP address of the broker.
• bundle: bundle=${bundle}. ${bundle} is the bundle range on this broker.
• metric: metric="assign".

Name	Type	Description
pulsar_lb_assign_broker_breakdown_total{result, reason}*	Counter	Assign broker breakdown count grouped by result and reason labels.

Service unit state channel metrics

All the service unit state channel metrics are labeled with the following labels:

• cluster: cluster=${pulsar_cluster}. ${pulsar_cluster} is the cluster name you have configured in the broker.conf file.
• metric: metric="sunitStateChn".

Name	Type	Description
pulsar_sunit_state_chn_owner_lookup_total{result, state}*	Counter	The owner broker lookup counts grouped by the result and state labels.
pulsar_sunit_state_chn_event_publish_ops_total{result, event}*	Counter	The published message count of service unit (e.g., bundle) state changes grouped by the result and event labels
pulsar_sunit_state_chn_subscribe_ops_total{result, event}*	Counter	The subscribed message count of service unit (e.g., bundle) state changes grouped by the result and event labels.
pulsar_sunit_state_chn_inactive_broker_cleanup_ops_total{result}*	Counter	The counts of inactive broker cleanup operations grouped by the result label.
pulsar_sunit_state_chn_orphan_su_cleanup_ops_total*	Counter	The total count of orphan service unit (e.g., bundle) cleanup operations.
pulsar_sunit_state_chn_owned_su_total*	Gauge	The number of owned bundles.
pulsar_sunit_state_chn_su_tombstone_cleanup_ops_total*	Counter	The total count of deleted service units (e.g., bundles) tombstone operations.
pulsar_sunit_state_chn_cleanup_ops_total{result=Failure}*	Counter	The total count of cleanup operation failures.

Additional load-manager diagnostics

These series depend on the configured load manager. Extensible load-manager and service-unit-state values below are exported as gauges, including fields whose names end in _total.

Name	Type	Description
pulsar_lb_ignored_ack_total	Gauge	Extensible load-manager ignored-ack diagnostic count.
pulsar_lb_ignored_send_total	Gauge	Extensible load-manager ignored-send diagnostic count.
pulsar_sunit_state_chn_metadata_state	Gauge	Service-unit channel metadata connection state: 0 (Stable), 1 (Jittery), or 2 (Unstable).
pulsar_sunit_state_chn_last_metadata_session_event_is_reestablished	Gauge	1 if the last metadata session event was session reestablishment; otherwise 0.
pulsar_sunit_state_chn_last_metadata_session_event_timestamp_ms	Gauge	Timestamp of the last metadata session event, in epoch milliseconds.
pulsar_sunit_state_chn_last_metadata_session_event_age_seconds	Gauge	Seconds since the last metadata session event, or -1 before an event has been recorded.
pulsar_lb_load_rank	Gauge	Simple load-manager broker load rank.
pulsar_lb_quota_pct_bandwidth_in	Gauge	Simple load-manager bandwidth in quota utilization percentage.
pulsar_lb_quota_pct_bandwidth_out	Gauge	Simple load-manager bandwidth out quota utilization percentage.
pulsar_lb_quota_pct_cpu	Gauge	Simple load-manager cpu quota utilization percentage.
pulsar_lb_quota_pct_memory	Gauge	Simple load-manager memory quota utilization percentage.

Extensible load-manager latency histograms

These collectors also retain brk_. They measure milliseconds and expose _bucket, _count, and _sum samples, with broker and metric labels. Buckets are 1, 10, 100, 200, 1,000 ms, and +Inf.

Name	Type	Description
brk_lb_unload_latency_ms	Histogram	Total unload duration on the source broker.
brk_lb_assign_latency_ms	Histogram	Time in the assignment state on the destination broker.
brk_lb_release_latency_ms	Histogram	Time in the release state on the source broker.
brk_lb_disconnect_latency_ms	Histogram	Time in the disconnected state on the source broker.

Pulsar Functions

Function instance metrics use tenant, namespace, name, instance_id, cluster, and fqfn (fully qualified function name). Worker metrics have their own labels, described below. The main instance table covers Java and Python names; see Go Functions metric names for runtime differences.

The following labels identify the cluster and namespace:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you have configured in the broker.conf file.
• namespace: namespace=\$\{pulsar_namespace\}. \$\{pulsar_namespace\} is the namespace name.

Name	Type	Description
pulsar_function_processed_successfully_total	Counter	The total number of messages processed successfully.
pulsar_function_processed_successfully_1min_total	Counter	The total number of messages processed successfully in the last 1 minute.
pulsar_function_system_exceptions_total	Counter	The total number of system exceptions.
pulsar_function_system_exceptions_1min_total	Counter	The total number of system exceptions in the last 1 minute.
pulsar_function_user_exceptions_total	Counter	The total number of user exceptions.
pulsar_function_user_exceptions_1min_total	Counter	The total number of user exceptions in the last 1 minute.
pulsar_function_process_latency_ms	Summary	The process latency in milliseconds.
pulsar_function_process_latency_ms_1min	Summary	The process latency in milliseconds in the last 1 minute.
pulsar_function_last_invocation	Gauge	The timestamp of the last invocation of the function.
pulsar_function_received_total	Counter	The total number of messages received from source.
pulsar_function_received_1min_total	Counter	The total number of messages received from source in the last 1 minute.
pulsar_function_user_metric_	Summary	The user-defined metrics.

Function exception details

Exception gauges include the exception text in the error label in addition to the instance labels.

Name	Type	Description
pulsar_function_source_exception	Gauge	Exception from source.
pulsar_function_sink_exception	Gauge	Exception from sink.
pulsar_function_user_exception	Gauge	Exception from user code, with an error label.
pulsar_function_system_exception	Gauge	Exception from system code, with an error label.

Go Functions metric names

The Go runtime uses the following names for its one-minute gauges. Java and Python use the _1min_total names in the main Functions table. These per-minute values reset and should not be treated as lifetime counters.

Name	Type	Description
pulsar_function_processed_successfully_total_1min	Gauge	Messages processed successfully in the current minute.
pulsar_function_system_exceptions_total_1min	Gauge	System exceptions in the current minute.
pulsar_function_user_exceptions_total_1min	Gauge	User exceptions in the current minute.
pulsar_function_received_total_1min	Gauge	Messages received in the current minute.
pulsar_function_user_metric	Summary	User-defined observations, distinguished by the metric label.

Functions worker metrics

Worker metrics use the cluster label. Leader-only metrics are emitted by the worker that currently leads scheduling.

Name	Type	Description
pulsar_function_worker_drain_execution_time_total_ms	Summary	Total execution time of a drain in milliseconds.
pulsar_function_worker_instance_count	Gauge	Number of instances run by this worker.
pulsar_function_worker_rebalance_execution_time_total_ms	Summary	Total execution time of a rebalance in milliseconds.
pulsar_function_worker_rebalance_strategy_execution_time_ms	Summary	Execution time of rebalance strategy in milliseconds.
pulsar_function_worker_schedule_execution_time_total_ms	Summary	Total execution time of schedule in milliseconds.
pulsar_function_worker_schedule_strategy_execution_time_ms	Summary	Execution time of schedule strategy in milliseconds.
pulsar_function_worker_start_instance_process_time_ms	Summary	Starting instance process time in milliseconds.
pulsar_function_worker_start_up_time_ms	Summary	Worker service startup time in milliseconds.
pulsar_function_worker_stop_instance_process_time_ms	Summary	Stopping instance process time in milliseconds.
pulsar_function_worker_total_function_count	Gauge	Total registered Functions known to the leader.
pulsar_function_worker_total_expected_instance_count	Gauge	Total expected Function instances known to the leader.
pulsar_function_worker_is_leader	Gauge	1 on the leader worker; emitted by the leader.

Connectors

All the Pulsar connector metrics are labeled with the following labels:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you have configured in the broker.conf file.
• namespace: namespace=\$\{pulsar_namespace\}. \$\{pulsar_namespace\} is the namespace name.

Connector metrics contain source metrics and sink metrics.

• Source metrics

  Name	Type	Description
  pulsar_source_written_total	Counter	The total number of records written to a Pulsar topic.
  pulsar_source_written_1min_total	Counter	The total number of records written to a Pulsar topic in the last 1 minute.
  pulsar_source_received_total	Counter	The total number of records received from source.
  pulsar_source_received_1min_total	Counter	The total number of records received from source in the last 1 minute.
  pulsar_source_last_invocation	Gauge	The timestamp of the last invocation of the source.
  pulsar_source_source_exception	Gauge	The exception from a source.
  pulsar_source_source_exceptions_total	Counter	The total number of source exceptions.
  pulsar_source_source_exceptions_1min_total	Counter	The total number of source exceptions in the last 1 minute.
  pulsar_source_system_exception	Gauge	The exception from system code.
  pulsar_source_system_exceptions_total	Counter	The total number of system exceptions.
  pulsar_source_system_exceptions_1min_total	Counter	The total number of system exceptions in the last 1 minute.
  pulsar_source_user_metric_	Summary	The user-defined metrics.

• Sink metrics

  Name	Type	Description
  pulsar_sink_written_total	Counter	The total number of records processed by a sink.
  pulsar_sink_written_1min_total	Counter	The total number of records processed by a sink in the last 1 minute.
  pulsar_sink_received_1min_total	Counter	The total number of messages that a sink has received from Pulsar topics in the last 1 minute.
  pulsar_sink_received_total	Counter	The total number of records that a sink has received from Pulsar topics.
  pulsar_sink_last_invocation	Gauge	The timestamp of the last invocation of the sink.
  pulsar_sink_sink_exception	Gauge	The exception from a sink.
  pulsar_sink_sink_exceptions_total	Counter	The total number of sink exceptions.
  pulsar_sink_sink_exceptions_1min_total	Counter	The total number of sink exceptions in the last 1 minute.
  pulsar_sink_system_exception	Gauge	The exception from system code.
  pulsar_sink_system_exceptions_total	Counter	The total number of system exceptions.
  pulsar_sink_system_exceptions_1min_total	Counter	The total number of system exceptions in the last 1 minute.
  pulsar_sink_user_metric_	Summary	The user-defined metrics.

Proxy

All the proxy metrics are labeled with the following labels:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you have configured in the broker.conf file.
• kubernetes_pod_name: kubernetes_pod_name=\$\{kubernetes_pod_name\}. \$\{kubernetes_pod_name\} is the Kubernetes pod name.

Name	Type	Description
pulsar_proxy_active_connections	Gauge	Number of connections currently active in the proxy.
pulsar_proxy_new_connections	Counter	Counter of connections being opened in the proxy.
pulsar_proxy_rejected_connections	Counter	Counter for connections rejected due to throttling.
pulsar_proxy_binary_ops	Counter	Counter of proxy operations.
pulsar_proxy_binary_bytes	Counter	Counter of proxy bytes.

Proxy lookup requests

These counters track proxy requests and throttling rejections.

Name	Type	Description
pulsar_proxy_get_schema_requests_total	Counter	Counter of schema requests
pulsar_proxy_get_topics_of_namespace_requests_total	Counter	Counter of getTopicsOfNamespace requests
pulsar_proxy_lookup_requests_total	Counter	Counter of topic lookup requests
pulsar_proxy_partitions_metadata_requests_total	Counter	Counter of partitions metadata requests
pulsar_proxy_rejected_get_topics_of_namespace_requests_total	Counter	Counter of getTopicsOfNamespace requests rejected due to throttling
pulsar_proxy_rejected_lookup_requests_total	Counter	Counter of topic lookup requests rejected due to throttling
pulsar_proxy_rejected_partitions_metadata_requests_total	Counter	Counter of partitions metadata requests rejected due to throttling

Pulsar transaction

All the transaction metrics are labeled with the following labels:

• cluster: cluster=\$\{pulsar_cluster\}. \$\{pulsar_cluster\} is the cluster name that you have configured in the broker.conf file.
• coordinator_id: coordinator_id=\$\{coordinator_id\}. \$\{coordinator_id\} is the coordinator id.

Name	Type	Description
pulsar_txn_active_count	Gauge	Number of active transactions.
pulsar_txn_created_total	Counter	Number of created transactions.
pulsar_txn_committed_total	Counter	Number of committed transactions.
pulsar_txn_aborted_total	Counter	Number of aborted transactions of this coordinator.
pulsar_txn_timeout_total	Counter	Number of timeout transactions.
pulsar_txn_append_log_total	Counter	Number of append transaction logs.
pulsar_txn_execution_latency_le_*	Histogram	Transaction execution latency. Available latencies are as below: latency="10" is TransactionExecutionLatency between (0ms, 10ms] latency="20" is TransactionExecutionLatency between (10ms, 20ms] latency="50" is TransactionExecutionLatency between (20ms, 50ms] latency="100" is TransactionExecutionLatency between (50ms, 100ms] latency="500" is TransactionExecutionLatency between (100ms, 500ms] latency="1000" is TransactionExecutionLatency between (500ms, 1000ms] latency="5000" is TransactionExecutionLatency between (1s, 5s] latency="15000" is TransactionExecutionLatency between (5s, 15s] latency="30000" is TransactionExecutionLatency between (15s, 30s] latency="60000" is TransactionExecutionLatency between (30s, 60s] latency="300000" is TransactionExecutionLatency between (1m,5m] latency="1500000" is TransactionExecutionLatency between (5m,15m] latency="3000000" is TransactionExecutionLatency between (15m,30m] latency="overflow" is TransactionExecutionLatency between (30m,∞]
pulsar_txn_tb_client_abort_failed_total	Counter	The number of failures to abort transactions for transaction buffer client.
pulsar_txn_tb_client_commit_failed_total	Counter	The number of failures to commit transaction for transaction buffer client.
pulsar_txn_tb_client_abort_latency	Summary	The latency of aborting transactions for transaction buffer client.
pulsar_txn_tb_client_commit_latency	Summary	The latency of committing transactions for transaction buffer client.
pulsar_txn_tb_client_pending_requests	Counter	The number of pending requests for transaction buffer client.
pulsar_txn_tp_committed_count_total	Counter	The number of committed transactions for pending ack store.
pulsar_txn_tp_aborted_count_total	Counter	The number of aborted transactions for pending ack store.
pulsar_txn_tp_commit_latency	Summary	The latency of committing transactions for transaction pending ack handle.

Resource groups

Resource-group counters accumulate reported usage, computed quotas, and management operations. Usage counters are cumulative additions of each interval's usage, not gauges for the most recent interval. Labels are ResourceGroup, plus MonitoringClass for usage/quota series and RemoteBroker for remote usage series. Registration counters are grouped by resource group rather than individual tenant or namespace.

Name	Type	Description
pulsar_resource_group_aggregate_usage_secs	Summary	Time required to aggregate usage of all resource groups, in seconds.
pulsar_resource_group_bytes_used_total	Counter	Bytes locally used within this resource group during the last aggregation interval
pulsar_resource_group_calculate_quota_secs	Summary	Time required to calculate quota of all resource groups, in seconds.
pulsar_resource_group_calculated_bytes_quota_total	Counter	Bytes quota calculated for resource group
pulsar_resource_group_calculated_messages_quota_total	Counter	Messages quota calculated for resource group
pulsar_resource_group_local_usage_reported_total	Counter	Number of times local usage was reported (vs. suppressed due to negligible change)
pulsar_resource_group_messages_used_total	Counter	Messages locally used within this resource group during the last aggregation interval
pulsar_resource_group_namespace_registers_total	Counter	Number of registrations of namespaces
pulsar_resource_group_namespace_unregisters_total	Counter	Number of un-registrations of namespaces
pulsar_resource_group_remote_usage_bytes_used_total	Counter	Bytes used reported about this resource group and monitoring class from a remote broker
pulsar_resource_group_remote_usage_messages_used_total	Counter	Messages used reported about this resource group and monitoring class from a remote broker
pulsar_resource_group_tenant_registers_total	Counter	Number of registrations of tenants
pulsar_resource_group_tenant_unregisters_total	Counter	Number of un-registrations of tenants
pulsar_resource_group_updates_total	Counter	Number of update operations on the given resource group

Topic-list memory limiting (broker and proxy)

For each suffix below, brokers expose pulsar_broker_topic_list_ plus the suffix, and proxies expose pulsar_proxy_topic_list_ plus the suffix. For example, pulsar_broker_topic_list_heap_memory_used_bytes and pulsar_proxy_topic_list_heap_memory_used_bytes measure heap usage. These metrics have no per-topic labels. Configure limits, queue capacities, and timeouts with the maxTopicListInFlightHeapMem* and maxTopicListInFlightDirectMem* settings.

Suffix	Type	Description
heap_memory_used_bytes	Gauge	Memory currently reserved by topic-list operations, in bytes.
heap_memory_limit_bytes	Gauge	Configured memory limit, in bytes.
heap_queue_size	Gauge	Requests waiting for memory permits.
heap_queue_max_size	Gauge	Configured maximum waiting queue size.
heap_wait_time_ms	Summary	Time waiting for memory permits, in milliseconds.
heap_timeout_total	Counter	Memory permit acquisition timeouts.
direct_memory_used_bytes	Gauge	Memory currently reserved by topic-list operations, in bytes.
direct_memory_limit_bytes	Gauge	Configured memory limit, in bytes.
direct_queue_size	Gauge	Requests waiting for memory permits.
direct_queue_max_size	Gauge	Configured maximum waiting queue size.
direct_wait_time_ms	Summary	Time waiting for memory permits, in milliseconds.
direct_timeout_total	Counter	Memory permit acquisition timeouts.