Skip to content
This repository has been archived by the owner on Jun 1, 2021. It is now read-only.

Latest commit

 

History

History
77 lines (57 loc) · 3.42 KB

README.md

File metadata and controls

77 lines (57 loc) · 3.42 KB
Raw

Eventuate replication endpoint metrics for dropwizard's metrics lib

eventuate-tool's dropwizard-metrics provides Gauges that returns metrics of an ReplicationEndpoint.

Project dependency

The following artifact is published to jfrog's snapshot and release repository:

  • Artifact Id: dropwizard-metrics_<scala-version>
  • Group Name: com.rbmhtechnology.eventuate-tools

Settings for an sbt-build:

libraryDependencies += "com.rbmhtechnology.eventuate-tools" %% "dropwizard-metrics" % "<version>"

// for snapshots
resolvers += "OJO Snapshots" at "https://oss.jfrog.org/oss-snapshot-local"

// for releases
resolvers += "OJO Releases" at "https://oss.jfrog.org/oss-release-local"

Start/Stop recording metrics

To register gauges for replication metrics and start recording metrics a DropwizardReplicationMetricsRecorder has to be initialized with the ReplicationEndpoint to be monitored and the MetricRegistry where gauges should be registered. A name prefix for all gauges can be optionally provided:

val endpoint = new ReplicationEndpoint(...)
val metricRegistry = new MetricRegistry
val recorder = new DropwizardReplicationMetricsRecorder(endpoint, metricRegistry, Some("prefix"))

Recording of metrics starts immediately and can be stopped with:

recorder.stopRecording()

Registered Gauges

Once recording is started the gauges for the following values are registered:

  • The sequence number of the local log with id under the name <prefix>.<log-id>.sequenceNo.
  • The time entry with id in the version vector of the local log with id under the name <prefix>.<log-id>.localVersionVector.<process-id>. For details see also the Eventuate documentation.
  • For each remote replica with id of the local log with id the replication progress in form of the sequence number in the remote log up to which events have been replicated to the local log under the name <prefix>.<log-id>.replicationProgress.<remote-id>

For checking the actual replication progress (i.e. how many events need to be replicated from one log to another for a full replication) the local sequence number of one log has to be compared with the replication progress of this replica on a remote node. For example in case of two replication endpoints (endpointA and endpointB) on two nodes (A and B) with a single replicated log (log1) and the following recorded values:

  • on node A
    • endpointA_log1.sequenceNo = 4
    • endpointA_log1.replicationProgress.endpointB_log1 = 3
  • on node B
    • endpointB_log1.sequenceNo = 5
    • endpointB_log1.replicationProgress.endpointA_log1 = 4

one can see that the replication from A to B has caught up with A's local progress as endpointA_log1.sequenceNo (4) equals endpointB_log1.replicationProgress.endpointA_log1 (4). As opposed to this the replication from B to A lacks behind as endpointB_log1.sequenceNo (5) is greater than endpointA_log1.replicationProgress.endpointB_log1 (3)