Elasticsearch OpenTelemetry metrics reference

This reference guide provides a comprehensive overview of all metrics collected by the Elasticsearch OpenTelemetry integration, helping you understand what data is available for monitoring your cluster health and performance.

Key metrics at a glance

Monitor your Elasticsearch cluster health and performance with these essential metrics:

Metric category	What it measures	Priority
Cluster health	`elasticsearch.cluster.health` - Overall cluster status (green/yellow/red)	🔴 Critical
Shard status	`elasticsearch.cluster.shards` - Unassigned, relocating, or initializing shards	🔴 Critical
Node availability	`elasticsearch.cluster.data_nodes` - Active data nodes in cluster	🔴 Critical
JVM heap usage	`jvm.memory.heap.utilization` - Memory usage percentage	🔴 Critical
Search performance	`elasticsearch.node.operations.time` - Query and fetch latency	🟡 Important
Resource usage	`system.cpu.utilization`, `system.memory.usage` - Host system resources	🔵 Monitoring

Complete metrics reference

The integration collects 50+ metrics across cluster, node, JVM, and host infrastructure. Expand the sections below for detailed metric specifications.

Tip

For the complete catalog of available metrics, see the OpenTelemetry elasticsearchreceiver and hostmetricsreceiver documentation.

Metric	Description	Attributes
`elasticsearch.cluster.health`	The health status of the cluster.	—
`elasticsearch.cluster.in_flight_fetch`	Shard fetch operations still in flight.	—
`elasticsearch.cluster.nodes`	Total cluster node count.	—
`elasticsearch.cluster.data_nodes`	The number of data nodes in the cluster.	—
`elasticsearch.cluster.shards`	The number of shards in the cluster.	—
`elasticsearch.cluster.pending_tasks`	Pending cluster-level tasks awaiting execution.	—
`elasticsearch.cluster.state_update.time`	Cumulative time spent updating cluster state.	`state` (any) `type` (computation, context_construction, commit, completion, master_apply, notification)

aggregation (primary_shards | total)

operation (index, delete, get, query, fetch, scroll, suggest, merge, refresh, flush, warmer)
aggregation (primary_shards | total)

Metric	Description	Attributes
`elasticsearch.index.documents`	Documents per index, split by state.	`state` (active \| deleted) `aggregation` (primary_shards \| total)
`elasticsearch.index.operations.merge.current`	Active segment merge operations.	`aggregation` (primary_shards \| total)
`elasticsearch.index.operations.time`	Time spent on index-level operations.	`operation` (index, delete, get, query, fetch, scroll, suggest, merge, refresh, flush, warmer) `aggregation` (primary_shards \| total)
`elasticsearch.index.shards.size`	The size of the shards assigned to this index.
`elasticsearch.index.segments.count`	Number of segments of an index.
`elasticsearch.index.operations.completed`	The number of operations completed for an index.

Metric	Description	Attributes
`elasticsearch.node.cache.count`	Query cache hits and misses across node shards.	`type` (hit \| miss)
`elasticsearch.node.cache.evictions`	Node cache evictions.	`cache_name` (fielddata \| query)
`elasticsearch.node.cache.memory.usage`	Cache memory usage in bytes.	`cache_name` (fielddata \| query)
`elasticsearch.node.cluster.io`	Internal cluster network I/O in bytes.	`direction` (received \| sent)
`elasticsearch.node.documents`	Documents hosted by the node.	`state` (active \| deleted)
`elasticsearch.node.disk.io.read`	Disk read throughput (KiB) across file stores.	—
`elasticsearch.node.disk.io.write`	Disk write throughput (KiB) across file stores.	—
`elasticsearch.node.fs.disk.available`	Disk available to the JVM.	—
`elasticsearch.node.fs.disk.total`	Total disk capacity on the node.	—
`elasticsearch.node.http.connections`	HTTP connections served by the node.	—
`elasticsearch.node.ingest.documents.current`	Documents currently being ingested.	—
`elasticsearch.node.ingest.operations.failed`	Cumulative ingest failures.	—
`elasticsearch.node.open_files`	Open file descriptors in use.	—
`elasticsearch.node.operations.completed`	Operations completed by the node.	`operation` (index, delete, get, query, fetch, scroll, suggest, merge, refresh, flush, warmer)
`elasticsearch.node.operations.current`	Operations currently in progress.	`operation` (same set)
`elasticsearch.node.operations.get.completed`	GET hits and misses.	`result` (hit \| miss)
`elasticsearch.node.operations.time`	Time spent on operations by a node..	`Operation` (index \| delete\| get\| query\| fetch\| scroll\| suggest\| merge\| refresh\| flush\| warmer)
`elasticsearch.node.shards.reserved.size`	Predicted shard growth due to recoveries.	—
`elasticsearch.node.shards.size`	The size of the shards assigned to this node.	—

Metric	Description	Attributes
`elasticsearch.os.cpu.load_avg.1m`	One-minute OS load average.	—
`elasticsearch.os.cpu.load_avg.5m`	Five-minute OS load average.	—
`elasticsearch.os.cpu.load_avg.15m`	Fifteen-minute OS load average.	—
`elasticsearch.os.cpu.usage`	Recent CPU usage for the whole system, or -1 if not supported.
`elasticsearch.os.memory`	Physical memory usage as seen by Elasticsearch.	`state` (free \| used)
`jvm.gc.collections.count`	Total garbage collection runs.	`name` (collector name)
`jvm.gc.collections.elapsed`	Time spent in garbage collection.	`name` (collector name)
`jvm.memory.heap.max`	Maximum heap memory available.	—
`jvm.memory.heap.used`	Heap memory currently in use.	—
`jvm.memory.heap.utilization`	Fraction of heap memory usage.	—
`jvm.threads.count`	Active JVM thread count.	—

Metric	Description	Attributes
`system.cpu.time`	Cumulative CPU time split by state.	`cpu` (logical CPU) `state` (idle\| interrupt\| nice\| softirq\| steal\| system\| user\| wait)
`system.cpu.utilization`	Difference in system.cpu.time since the last measurement per logical CPU, divided by the elapsed time (value in interval [0,1]).	`cpu` (logical CPU) `state` (same set as above)
`system.cpu.load_average.1m`	One-minute system load average.	—
`system.cpu.load_average.5m`	Five-minute system load average.	—
`system.cpu.load_average.15m`	Fifteen-minute system load average.	—
`system.memory.utilization`	Memory utilization ratio.	`state` (used)
`system.memory.usage`	Bytes of memory in use..	`state` (buffered\| cached\| inactive\| free\| slab_reclaimable\| slab_unreclaimable\| used)
`system.disk.io`	Disk I/O throughput per device.	`device` (disk) `direction` (read \| write)
`system.disk.operations`	Disk operations per device.	`device` `direction` (read \| write)
`system.filesystem.usage`	Filesystem capacity by state.	`device` `state` (used \| free \| reserved)
`system.filesystem.utilization`	Filesystem utilization ratio.	`device` `state` (used)
`system.network.io`	Network bytes per interface.	`interface` `direction` (receive \| transmit)
`system.network.packets`	Network packets per interface.	`interface` `direction` (receive \| transmit)
`process.cpu.utilization`	Percentage of total CPU time used by the process since last scrape, expressed as a value between 0 and 1.	`process.pid` (PID) `process.executable.name` (binary name) `process.owner` (user) `state` (user \| system \| other)

Resource attributes by deployment type

All Elasticsearch metrics include resource attributes (tags) that help you organize and filter your data in New Relic. The specific attributes depend on how you've deployed Elasticsearch:

Common attributes (all deployments)

Every deployment includes these core Elasticsearch identifiers:

Attribute	Description
`elasticsearch.cluster.name`	The unique name of your Elasticsearch cluster
`elasticsearch.node.name`	Individual Elasticsearch node identifier

Host-based deployments

For Elasticsearch running directly on hosts or VMs, you'll see additional host infrastructure attributes:

Attribute	Description
`host.name`	Hostname where Elasticsearch is running
`os.type`	Operating system (linux, windows, darwin)

Kubernetes deployments

Kubernetes deployments include additional container orchestration attributes for enhanced visibility:

Attribute	Description
`k8s.cluster.name`	Your Kubernetes cluster name
`k8s.namespace.name`	The namespace where Elasticsearch is running
`k8s.pod.name`	The specific Elasticsearch pod name
`k8s.pod.uid`	Unique identifier for the pod
`k8s.deployment.name`	Kubernetes deployment name (if using Deployment)

Using resource attributes effectively

These attributes enable powerful monitoring scenarios:

Multi-environment management:

Filter dashboards by elasticsearch.cluster.name for production vs staging
Create cluster-specific alert policies
Compare performance across different environments

Infrastructure correlation:

Use host.name or k8s.pod.name to correlate with infrastructure monitoring
Track resource utilization at the host or pod level
Identify performance patterns across your infrastructure

Kubernetes-specific monitoring:

Monitor Elasticsearch across multiple namespaces
Track pod lifecycle events and their performance impact
Create alerts based on Kubernetes deployment health

Next steps

Now that you understand what metrics are available, here are your next steps for effective Elasticsearch monitoring:

Installation and configuration:

Host installation - Set up monitoring for Elasticsearch running on hosts/VMs
Kubernetes installation - Configure monitoring for Elasticsearch running on Kubernetes

Set up monitoring and alerts:

View your data - Learn how to access dashboards and explore your Elasticsearch metrics in New Relic
Create alerts - Set up proactive monitoring with guided mode or NRQL alert conditions
NRQL queries - Write custom queries to analyze your Elasticsearch performance data

Elasticsearch OpenTelemetry metrics reference

Key metrics at a glance

Complete metrics reference

Tip

Cluster metrics(elasticsearchreceiver)

Index activity metrics (elasticsearchreceiver)

Node activity & workload metrics (elasticsearchreceiver)

JVM & OS metrics reported by Elasticsearch (elasticsearchreceiver)

Host infrastructure metrics (hostmetricsreceiver)

Resource attributes by deployment type

Common attributes (all deployments)

Host-based deployments

Kubernetes deployments

Using resource attributes effectively

Next steps

Elasticsearch OpenTelemetry metrics reference

Key metrics at a glance.css-21sua1{background:none;border:none;width:0;padding:0;}

Complete metrics reference

Tip

Index activity metrics (elasticsearchreceiver)

Node activity & workload metrics (elasticsearchreceiver)

JVM & OS metrics reported by Elasticsearch (elasticsearchreceiver)

Host infrastructure metrics (hostmetricsreceiver)

Resource attributes by deployment type

Common attributes (all deployments)

Host-based deployments

Kubernetes deployments

Using resource attributes effectively

Next steps

Key metrics at a glance