This reference guide provides a comprehensive overview of all metrics collected by the Redis OpenTelemetry integration, helping you understand what data is available for monitoring your Redis instance health and performance.
Key metrics at a glance
Monitor your Redis instance health and performance with these essential metrics:
Metric category
What it measures
Priority
Memory utilization
redis.memory.used, redis.maxmemory — Current memory consumption vs configured limit
Critical
Cache hit ratio
redis.keyspace.hits, redis.keyspace.misses — Ratio of successful key lookups
Critical
Key evictions
redis.keys.evicted — Keys removed due to memory pressure
Critical
Commands/sec
redis.commands, redis.commands.processed — Throughput of commands being executed
Important
Connected clients
redis.clients.connected, redis.clients.blocked — Active and waiting connections
Important
Fragmentation ratio
redis.memory.fragmentation_ratio — Ratio of OS-allocated to Redis-allocated memory
Monitoring
Cluster health
redis.cluster.state, redis.cluster.slots_ok, redis.cluster.known_nodes — Overall cluster status and slot coverage (Cluster mode only)
Critical
Complete metrics reference
The integration collects 30+ instance metrics across commands, memory, keyspace, connections, and replication — plus 9 cluster-specific metrics for Redis Cluster deployments. Expand the sections below for detailed metric specifications.
Number of client connections (excluding replicas).
—
redis.clients.max_input_buffer
Biggest input buffer among current client connections (bytes).
—
redis.clients.max_output_buffer
Longest output list among current client connections (bytes).
—
Metric
Description
Attributes
redis.commands
Number of commands processed per second (instantaneous).
—
redis.commands.processed
Total number of commands processed by the server.
—
redis.cmd.calls
Total number of calls for a command.
cmd
redis.cmd.latency
Command execution latency (seconds).
cmd, percentile
redis.cmd.usec
Total time for all executions of a command (microseconds).
cmd
Metric
Description
Attributes
redis.memory.used
Total bytes allocated by Redis using its allocator.
—
redis.memory.peak
Peak memory consumed by Redis (bytes).
—
redis.memory.rss
Number of bytes allocated as seen by the operating system.
—
redis.memory.lua
Number of bytes used by the Lua engine.
—
redis.memory.fragmentation_ratio
Ratio of OS-allocated memory to Redis-allocated memory.
—
redis.maxmemory
The value of the maxmemory configuration directive (bytes).
—
Metric
Description
Attributes
redis.keyspace.hits
Number of successful lookup of keys in the main dictionary.
—
redis.keyspace.misses
Number of failed lookup of keys in the main dictionary.
—
redis.keys.evicted
Number of evicted keys due to maxmemory limit.
—
redis.keys.expired
Total number of key expiration events.
—
redis.db.keys
Number of keyspace keys.
db
redis.db.expires
Number of keyspace keys with an expiration.
db
redis.db.avg_ttl
Average keyspace keys TTL (milliseconds).
db
Metric
Description
Attributes
redis.connections.received
Total number of connections accepted by the server.
—
redis.connections.rejected
Number of connections rejected because of maxclients limit.
—
redis.net.input
Total number of bytes read from the network.
—
redis.net.output
Total number of bytes written to the network.
—
Metric
Description
Attributes
redis.slaves.connected
Number of connected replicas.
—
redis.replication.offset
The server's current replication offset (bytes).
—
redis.replication.backlog_first_byte_offset
The master offset of the replication backlog buffer (bytes).
—
redis.role
Replication role of this Redis instance.
—
Metric
Description
Attributes
redis.uptime
Number of seconds since Redis server start.
—
redis.cpu.time
CPU time consumed by Redis (seconds).
state
redis.latest_fork
Duration of the latest fork operation (microseconds).
—
redis.rdb.changes_since_last_save
Number of changes since the last RDB dump.
—
Cluster metrics
These metrics are only available when monitoring Redis in Cluster mode. They require enabling cluster metrics in the Redis receiver (all cluster metrics are disabled by default) or using the Prometheus receiver approach with redis_exporter --is-cluster.
Metric
Description
Type
redis.cluster.state
Overall cluster health status. 1 = OK (all slots served), 0 = FAIL (one or more slots unavailable).
Gauge
redis.cluster.slots_assigned
Number of hash slots mapped to nodes. Must be 16384 for a fully operational cluster.
Gauge
redis.cluster.slots_ok
Number of slots actively serving requests without issues.
Gauge
redis.cluster.slots_pfail
Number of slots in PFAIL (probable failure) state — a node is suspected unreachable but not yet confirmed.
Gauge
redis.cluster.slots_fail
Number of slots in confirmed FAIL state — NOT serving traffic.
Gauge
redis.cluster.known_nodes
Total number of nodes (masters + replicas) visible in the cluster.
Gauge
redis.cluster.node.count
Number of master nodes currently serving hash slots.
Gauge
redis.cluster.stats_messages_sent
Total cluster bus messages sent (ping, pong, meet, fail). Steady traffic indicates healthy gossip protocol.
Cumulative
redis.cluster.stats_messages_received
Total cluster bus messages received from peer nodes. A sudden drop indicates possible network partition.
Cumulative
Conseil
All nodes in a Redis Cluster report identical cluster metrics (CLUSTER INFO returns the same data from any node). You can enable cluster monitoring on all nodes for resilience or on a single node to minimize redundant data. For the Prometheus receiver approach, use redis_exporter --is-cluster which automatically collects these metrics.
Resource attributes
Resource attributes are attached to all metrics and help you filter and group your data in New Relic.
Attribute
Description
server.address
Hostname or IP of the Redis server
server.port
Port number of the Redis server
redis.instance.id
Custom identifier for the Redis instance
redis.version
Redis server version
host.name
Hostname where the collector is running
host.id
Unique identifier of the host
k8s.cluster.name
Kubernetes cluster name
k8s.namespace.name
Kubernetes namespace
k8s.pod.name
Kubernetes pod name
redis.cluster.name
User-defined Redis Cluster name (only present on cluster metrics)