[HDDS-8746] Add metrics to ReplicationSupervisor for task count and max stream

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.4.0
• Component/s: SCM
• Labels: pull-request-available
• Assignee: Stephen O'Donnell
• Reporter: Stephen O'Donnell
• Created: 01/Jun/23 16:15
• Updated: 02/Jun/23 10:59
• Resolved: 02/Jun/23 10:59

Description

To give added visibility into the Replication Queue on a datanode, this Jira adds some extra metrics to give visibility into the count of "normal priority" Replication and Reconstruction tasks, along with the current maxStreams setting, which is automatically adjusted when the node moves into decommission.

This was tested manually in Docker compose by sampling the JMX output. Sample output:

    "name" : "Hadoop:service=HddsDatanode,name=ReplicationSupervisorMetrics",
    "modelerType" : "ReplicationSupervisorMetrics",
    "tag.Hostname" : "b297306fcf0f",
    "numInFlightReplications" : 5,
    "numQueuedReplications" : 4,
    "numRequestedReplications" : 1,
    "numTimeoutReplications" : 0,
    "numSkippedReplications" : 0,
    "maxReplicationStreams" : 1, ** newly added
    "numInflightECReconstructionCoordinatorTask" : 5 ** newly added


    "name" : "Hadoop:service=HddsDatanode,name=ReplicationSupervisorMetrics",
    "modelerType" : "ReplicationSupervisorMetrics",
    "tag.Hostname" : "b297306fcf0f",
    "numInFlightReplications" : 1,
    "numQueuedReplications" : 0,
    "numRequestedReplications" : 6,
    "numTimeoutReplications" : 0,
    "numSkippedReplications" : 0,
    "maxReplicationStreams" : 1,      ** newly added
    "numInflightReplicationTask" : 1,  ** newly added
    "numInflightECReconstructionCoordinatorTask" : 0 ** newly added

Issue Links

• links to: GitHub Pull Request #4818