etcd

Supported OS Linux Mac OS Windows

Integration version6.2.0

Etcd Dashboard

Overview

Collect Etcd metrics to:

  • Monitor the health of your Etcd cluster.
  • Know when host configurations may be out of sync.
  • Correlate the performance of Etcd with the rest of your applications.

Setup

Installation

The Etcd check is included in the Datadog Agent package, so you don’t need to install anything else on your Etcd instance(s).

Configuration

    Host

    To configure this check for an Agent running on a host:

    Metric collection
    1. Edit the etcd.d/conf.yaml file, in the conf.d/ folder at the root of your Agent’s configuration directory to start collecting your Etcd performance data. See the sample etcd.d/conf.yaml for all available configuration options.
    2. Restart the Agent
    Log collection

    1. Collecting logs is disabled by default in the Datadog Agent, enable it in your datadog.yaml file:

      logs_enabled: true

    2. Uncomment and edit this configuration block at the bottom of your etcd.d/conf.yaml:

      logs:
  - type: file
    path: "<LOG_FILE_PATH>"
    source: etcd
    service: "<SERVICE_NAME>"

      Change the path and service parameter values based on your environment. See the sample etcd.d/conf.yaml for all available configuration options.

    3. Restart the Agent.

    Containerized

    For containerized environments, see the Autodiscovery Integration Templates for guidance on applying the parameters below.

    Metric collection
    ParameterValue
    <INTEGRATION_NAME>etcd
    <INIT_CONFIG>blank or {}
    <INSTANCE_CONFIG>{"prometheus_url": "http://%%host%%:2379/metrics"}
    Log collection

    Collecting logs is disabled by default in the Datadog Agent. To enable it, see Kubernetes log collection.

    ParameterValue
    <LOG_CONFIG>{"source": "etcd", "service": "<SERVICE_NAME>"}

    Validation

    Run the Agent’s status subcommand and look for etcd under the Checks section.

    Data Collected

    Metrics

    etcd.debugging.mvcc.db.compaction.keys.total
    (count)    		Total number of db keys compacted.
    Shown as key
    etcd.debugging.mvcc.db.compaction.pause.duration.milliseconds
    (gauge)    		Bucketed histogram of db compaction pause duration.
    Shown as millisecond
    etcd.debugging.mvcc.db.compaction.total.duration.milliseconds
    (gauge)    		Bucketed histogram of db compaction total duration.
    Shown as millisecond
    etcd.debugging.mvcc.db.total.size.in_bytes
    (gauge)    		Total size of the underlying database in bytes.
    Shown as byte
    etcd.debugging.mvcc.delete.total
    (count)    		Total number of deletes seen by this member.
    Shown as query
    etcd.debugging.mvcc.events.total
    (count)    		Total number of events sent by this member.
    Shown as event
    etcd.debugging.mvcc.index.compaction.pause.duration.milliseconds
    (gauge)    		Bucketed histogram of index compaction pause duration.
    Shown as millisecond
    etcd.debugging.mvcc.keys.total
    (gauge)    		Total number of keys.
    Shown as key
    etcd.debugging.mvcc.pending.events.total
    (gauge)    		Total number of pending events to be sent.
    Shown as event
    etcd.debugging.mvcc.put.total
    (count)    		Total number of puts seen by this member.
    Shown as query
    etcd.debugging.mvcc.range.total
    (count)    		Total number of ranges seen by this member.
    Shown as query
    etcd.debugging.mvcc.slow_watcher.total
    (gauge)    		Total number of unsynced slow watchers.
    Shown as connection
    etcd.debugging.mvcc.txn.total
    (count)    		Total number of txns seen by this member.
    Shown as transaction
    etcd.debugging.mvcc.watch_stream.total
    (gauge)    		Total number of watch streams.
    Shown as connection
    etcd.debugging.mvcc.watcher.total
    (gauge)    		Total number of watchers.
    Shown as connection
    etcd.debugging.server.lease.expired.total
    (count)    		The total number of expired leases.
    Shown as item
    etcd.debugging.snap.save.marshalling.duration.seconds
    (gauge)    		The marshalling cost distributions of save called by snapshot.
    Shown as second
    etcd.debugging.snap.save.total.duration.seconds
    (gauge)    		The total latency distributions of save called by snapshot.
    Shown as second
    etcd.debugging.store.expires.total
    (count)    		Total number of expired keys.
    Shown as key
    etcd.debugging.store.reads.total
    (count)    		Total number of reads action by (get/getRecursive), local to this member.
    Shown as read
    etcd.debugging.store.watch.requests.total
    (count)    		Total number of incoming watch requests (new or reestablished).
    Shown as request
    etcd.debugging.store.watchers
    (gauge)    		Count of currently active watchers.
    Shown as connection
    etcd.debugging.store.writes.total
    (count)    		Total number of writes (e.g. set/compareAndDelete) seen by this member.
    Shown as write
    etcd.disk.backend.commit.duration.seconds
    (gauge)    		The latency distributions of commit called by backend.
    Shown as second
    etcd.disk.backend.snapshot.duration.seconds
    (gauge)    		The latency distribution of backend snapshots.
    Shown as second
    etcd.disk.wal.fsync.duration.seconds.count
    (count)    		The count of latency distributions of fsync called by wal.
    Shown as second
    etcd.disk.wal.fsync.duration.seconds.sum
    (gauge)    		The sum of latency distributions of fsync called by wal.
    Shown as second
    etcd.disk.wal.write.bytes.total
    (gauge)    		Total number of bytes written in WAL
    Shown as byte
    etcd.etcd.server.client.requests.total
    (count)    		The total number of client requests per client version
    Shown as request
    etcd.go.gc.duration.seconds
    (gauge)    		A summary of the GC invocation durations.
    Shown as second
    etcd.go.goroutines
    (gauge)    		Number of goroutines that currently exist.
    Shown as thread
    etcd.go.info
    (gauge)    		Information about the Go environment.
    Shown as item
    etcd.go.memstats.alloc.bytes
    (gauge)    		Number of bytes allocated and still in use.
    Shown as byte
    etcd.go.memstats.alloc.bytes.total
    (count)    		Total number of bytes allocated, even if freed.
    Shown as byte
    etcd.go.memstats.buck.hash.sys.bytes
    (gauge)    		Number of bytes used by the profiling bucket hash table.
    Shown as byte
    etcd.go.memstats.frees.total
    (count)    		Total number of frees.
    Shown as occurrence
    etcd.go.memstats.gc.cpu.fraction
    (gauge)    		The fraction of this program's available CPU time used by the GC since the program started.
    Shown as cpu
    etcd.go.memstats.gc.sys.bytes
    (gauge)    		Number of bytes used for garbage collection system metadata.
    Shown as byte
    etcd.go.memstats.heap.alloc.bytes
    (gauge)    		Number of heap bytes allocated and still in use.
    Shown as byte
    etcd.go.memstats.heap.idle.bytes
    (gauge)    		Number of heap bytes waiting to be used.
    Shown as byte
    etcd.go.memstats.heap.inuse.bytes
    (gauge)    		Number of heap bytes that are in use.
    Shown as byte
    etcd.go.memstats.heap.objects
    (gauge)    		Number of allocated objects.
    Shown as item
    etcd.go.memstats.heap.released.bytes
    (gauge)    		Number of heap bytes released to OS.
    Shown as byte
    etcd.go.memstats.heap.sys.bytes
    (gauge)    		Number of heap bytes obtained from system.
    Shown as byte
    etcd.go.memstats.last.gc.time.seconds
    (gauge)    		Number of seconds since 1970 of last garbage collection.
    Shown as second
    etcd.go.memstats.lookups.total
    (count)    		Total number of pointer lookups.
    Shown as occurrence
    etcd.go.memstats.mallocs.total
    (count)    		Total number of mallocs.
    Shown as occurrence
    etcd.go.memstats.mcache.inuse.bytes
    (gauge)    		Number of bytes in use by mcache structures.
    Shown as byte
    etcd.go.memstats.mcache.sys.bytes
    (gauge)    		Number of bytes used for mcache structures obtained from system.
    Shown as byte
    etcd.go.memstats.mspan.inuse.bytes
    (gauge)    		Number of bytes in use by mspan structures.
    Shown as byte
    etcd.go.memstats.mspan.sys.bytes
    (gauge)    		Number of bytes used for mspan structures obtained from system.
    Shown as byte
    etcd.go.memstats.next.gc.bytes
    (gauge)    		Number of heap bytes when next garbage collection will take place.
    Shown as byte
    etcd.go.memstats.other.sys.bytes
    (gauge)    		Number of bytes used for other system allocations.
    Shown as byte
    etcd.go.memstats.stack.inuse.bytes
    (gauge)    		Number of bytes in use by the stack allocator.
    Shown as byte
    etcd.go.memstats.stack.sys.bytes
    (gauge)    		Number of bytes obtained from system for stack allocator.
    Shown as byte
    etcd.go.memstats.sys.bytes
    (gauge)    		Number of bytes obtained from system.
    Shown as byte
    etcd.go.threads
    (gauge)    		Number of OS threads created.
    Shown as thread
    etcd.grpc.proxy.cache.hits.total
    (gauge)    		Total number of cache hits
    Shown as occurrence
    etcd.grpc.proxy.cache.keys.total
    (gauge)    		Total number of keys/ranges cached
    Shown as item
    etcd.grpc.proxy.cache.misses.total
    (gauge)    		Total number of cache misses
    Shown as occurrence
    etcd.grpc.proxy.events.coalescing.total
    (count)    		Total number of events coalescing
    Shown as event
    etcd.grpc.proxy.watchers.coalescing.total
    (gauge)    		Total number of current watchers coalescing
    Shown as connection
    etcd.grpc.server.handled.total
    (count)    		Total number of RPCs completed on the server, regardless of success or failure.
    Shown as operation
    etcd.grpc.server.msg.received.total
    (count)    		Total number of RPC stream messages received on the server.
    Shown as operation
    etcd.grpc.server.msg.sent.total
    (count)    		Total number of gRPC stream messages sent by the server.
    Shown as operation
    etcd.grpc.server.started.total
    (count)    		Total number of RPCs started on the server.
    Shown as operation
    etcd.leader.counts.fail
    (gauge)    		Rate of failed Raft RPC requests (ETCD API V2 only)
    Shown as request
    etcd.leader.counts.success
    (gauge)    		Rate of successful Raft RPC requests (ETCD API V2 only)
    Shown as request
    etcd.leader.latency.avg
    (gauge)    		Average latency to each peer in the cluster (ETCD API V2 only)
    Shown as millisecond
    etcd.leader.latency.current
    (gauge)    		Current latency to each peer in the cluster (ETCD API V2 only)
    Shown as millisecond
    etcd.leader.latency.max
    (gauge)    		Maximum latency to each peer in the cluster (ETCD API V2 only)
    Shown as millisecond
    etcd.leader.latency.min
    (gauge)    		Minimum latency to each peer in the cluster (ETCD API V2 only)
    Shown as millisecond
    etcd.leader.latency.stddev
    (gauge)    		Standard deviation latency to each peer in the cluster (ETCD API V2 only)
    Shown as millisecond
    etcd.mvcc.db.total.size.in_use.bytes
    (gauge)    		Total size of the underlying database logically in use
    Shown as byte
    etcd.network.active_peers
    (gauge)    		The current number of active peer connections
    Shown as connection
    etcd.network.client.grpc.received.bytes.total
    (count)    		The total number of bytes received from grpc clients.
    Shown as byte
    etcd.network.client.grpc.sent.bytes.total
    (count)    		The total number of bytes sent to grpc clients.
    Shown as byte
    etcd.network.disconnected_peers.total
    (count)    		The total number of disconnected peers
    Shown as connection
    etcd.network.peer.received.bytes.total
    (count)    		The total number of bytes received from peers.
    Shown as byte
    etcd.network.peer.received.failures.total
    (count)    		The total number of receive failures from peers
    Shown as event
    etcd.network.peer.round_trip_time.seconds
    (gauge)    		Round-Trip-Time histogram between peers.
    Shown as second
    etcd.network.peer.sent.bytes.total
    (count)    		The total number of bytes sent to peers.
    Shown as byte
    etcd.network.peer.sent.failures.total
    (count)    		The total number of send failures from peers
    Shown as event
    etcd.network.snapshot.receive.failures.total
    (count)    		Total number of snapshot receive failures
    Shown as event
    etcd.network.snapshot.receive.inflights.total
    (gauge)    		Total number of inflight snapshot sends
    Shown as event
    etcd.network.snapshot.receive.success.total
    (count)    		Total number of successful snapshot receives
    Shown as event
    etcd.network.snapshot.receive.total.duration.seconds.count
    (gauge)    		Total latency distributions of v3 snapshot receives
    Shown as second
    etcd.network.snapshot.receive.total.duration.seconds.sum
    (gauge)    		Total latency distributions of v3 snapshot receives
    Shown as second
    etcd.network.snapshot.send.failures.total
    (count)    		The total number of send failures from peers
    Shown as event
    etcd.network.snapshot.send.inflights.total
    (gauge)    		Total number of inflight snapshot receives
    Shown as event
    etcd.network.snapshot.send.sucess.total
    (count)    		Total number of successful snapshot sends
    Shown as event
    etcd.network.snapshot.send.total.duration.seconds.count
    (gauge)    		Total latency distributions of v3 snapshot sends
    Shown as second
    etcd.network.snapshot.send.total.duration.seconds.sum
    (gauge)    		Total latency distributions of v3 snapshot sends
    Shown as second
    etcd.os.fd.limit
    (gauge)    		The file descriptor limit
    Shown as object
    etcd.os.fd.used
    (gauge)    		The number of used file descriptors
    Shown as object
    etcd.process.cpu.seconds.total
    (count)    		Total user and system CPU time spent in seconds.
    Shown as cpu
    etcd.process.max.fds
    (gauge)    		Maximum number of open file descriptors.
    Shown as item
    etcd.process.open.fds
    (gauge)    		Number of open file descriptors.
    Shown as item
    etcd.process.resident.memory.bytes
    (gauge)    		Resident memory size in bytes.
    Shown as byte
    etcd.process.start.time.seconds
    (gauge)    		Start time of the process since unix epoch in seconds.
    Shown as second
    etcd.process.virtual.memory.bytes
    (gauge)    		Virtual memory size in bytes.
    Shown as byte
    etcd.self.recv.appendrequest.count
    (gauge)    		Rate of append requests this node has processed (ETCD API V2 only)
    Shown as request
    etcd.self.recv.bandwidthrate
    (gauge)    		Rate of bytes received (ETCD API V2 only)
    Shown as byte
    etcd.self.recv.pkgrate
    (gauge)    		Rate of packets received (ETCD API V2 only)
    Shown as packet
    etcd.self.send.appendrequest.count
    (gauge)    		Rate of append requests this node has sent (ETCD API V2 only)
    Shown as request
    etcd.self.send.bandwidthrate
    (gauge)    		Rate of bytes sent (ETCD API V2 only)
    Shown as byte
    etcd.self.send.pkgrate
    (gauge)    		Rate of packets sent (ETCD API V2 only)
    Shown as packet
    etcd.server.apply.slow.total
    (count)    		The total number of slow apply requests (likely overloaded from slow disk)
    Shown as request
    etcd.server.go_version
    (gauge)    		Which Go version server is running with. 1 with label with current version
    Shown as unit
    etcd.server.has_leader
    (gauge)    		Whether or not a leader exists. 1 is existence, 0 is not.
    Shown as check
    etcd.server.health.failures.total
    (count)    		The total number of failed health checks
    Shown as event
    etcd.server.health.success.total
    (count)    		The total number of successful health checks
    Shown as event
    etcd.server.heartbeat.send.failures.total
    (count)    		The total number of leader heartbeat send failures (likely overloaded from slow disk)
    Shown as event
    etcd.server.is_leader
    (gauge)    		Whether or not this member is a leader. 1 if is, 0 otherwise.
    Shown as check
    etcd.server.leader.changes.seen.total
    (count)    		The number of leader changes seen.
    Shown as event
    etcd.server.lease.expired.total
    (count)    		The total number of expired leases
    Shown as occurrence
    etcd.server.proposals.applied.total
    (gauge)    		The total number of consensus proposals applied.
    Shown as occurrence
    etcd.server.proposals.committed.total
    (gauge)    		The total number of consensus proposals committed.
    Shown as occurrence
    etcd.server.proposals.failed.total
    (count)    		The total number of failed proposals seen.
    Shown as occurrence
    etcd.server.proposals.pending
    (gauge)    		The current number of pending proposals to commit.
    Shown as occurrence
    etcd.server.quota.backend.bytes
    (gauge)    		Current backend storage quota size in bytes
    Shown as byte
    etcd.server.read_indexes.failed.total
    (count)    		The total number of failed read indexes seen
    Shown as event
    etcd.server.read_indexes.slow.total
    (count)    		The total number of pending read indexes not in sync with leader or timed out read index requests
    Shown as event
    etcd.server.version
    (gauge)    		Which version is running. 1 for 'server_version' label with current version.
    Shown as item
    etcd.snap.db.fsync.duration.seconds.count
    (gauge)    		The latency distributions of fsyncing .snap.db file
    Shown as second
    etcd.snap.db.fsync.duration.seconds.sum
    (gauge)    		The latency distributions of fsyncing .snap.db file
    Shown as second
    etcd.snap.db.save.total.duration.seconds.count
    (gauge)    		The total latency distributions of v3 snapshot save
    Shown as second
    etcd.snap.db.save.total.duration.seconds.sum
    (gauge)    		The total latency distributions of v3 snapshot save
    Shown as second
    etcd.snap.fsync.duration.seconds.count
    (gauge)    		The latency distributions of fsync called by snap
    Shown as second
    etcd.snap.fsync.duration.seconds.sum
    (gauge)    		The latency distributions of fsync called by snap
    Shown as second
    etcd.store.compareanddelete.fail
    (gauge)    		Rate of compare and delete requests failure (ETCD API V2 only)
    Shown as request
    etcd.store.compareanddelete.success
    (gauge)    		Rate of compare and delete requests success (ETCD API V2 only)
    Shown as request
    etcd.store.compareandswap.fail
    (gauge)    		Rate of compare and swap requests failure (ETCD API V2 only)
    Shown as request
    etcd.store.compareandswap.success
    (gauge)    		Rate of compare and swap requests success (ETCD API V2 only)
    Shown as request
    etcd.store.create.fail
    (gauge)    		Rate of failed create requests (ETCD API V2 only)
    Shown as request
    etcd.store.create.success
    (gauge)    		Rate of successful create requests (ETCD API V2 only)
    Shown as request
    etcd.store.delete.fail
    (gauge)    		Rate of failed delete requests (ETCD API V2 only)
    Shown as request
    etcd.store.delete.success
    (gauge)    		Rate of successful delete requests (ETCD API V2 only)
    Shown as request
    etcd.store.expire.count
    (gauge)    		Rate of expired keys (ETCD API V2 only)
    Shown as eviction
    etcd.store.gets.fail
    (gauge)    		Rate of failed get requests (ETCD API V2 only)
    Shown as request
    etcd.store.gets.success
    (gauge)    		Rate of successful get requests (ETCD API V2 only)
    Shown as request
    etcd.store.sets.fail
    (gauge)    		Rate of failed set requests (ETCD API V2 only)
    Shown as request
    etcd.store.sets.success
    (gauge)    		Rate of successful set requests (ETCD API V2 only)
    Shown as request
    etcd.store.update.fail
    (gauge)    		Rate of failed update requests (ETCD API V2 only)
    Shown as request
    etcd.store.update.success
    (gauge)    		Rate of successful update requests (ETCD API V2 only)
    Shown as request
    etcd.store.watchers
    (gauge)    		Rate of watchers(ETCD API V2 only)

    Etcd metrics are tagged with etcd_state:leader or etcd_state:follower, depending on the node status, so you can easily aggregate metrics by status.

    Events

    The Etcd check does not include any events.

    Service Checks

    etcd.can_connect
    Returns CRITICAL if unable to get metrics from etcd (timeout or non-200 HTTP code). This service check is only available on the legacy version of the etcd check.
    Statuses: ok, critical

    etcd.healthy
    Returns CRITICAL when a member is unhealthy. This service check is only available on the legacy version of the etcd check.
    Statuses: ok, critical, unknown

    etcd.prometheus.health
    Returns CRITICAL if the check cannot access a metrics endpoint. Otherwise, returns OK. This service check is only available when use_preview is enabled.
    Statuses: ok, critical

    Troubleshooting

    Need help? Contact Datadog support.

    Further Reading